AIA - Vision Online

Ask The Experts

Vision Systems Questions & Answers

Click a question to reveal the answer. Don't see the information you're looking for? Take a look at the archived questions or ask a new question by clicking the button to the right.

Investigating if a vision system exists or can be developed which can scan various electrical connection systems and determine if the connector is fully mated. The scanning device needs to be portable and self contained such that an operator can be free to move from vehicle to vehicle checking multiple connections and connector types at various angles and/or orientations.

Hi Bernard, It sounds a bit tricky because the environment (lighting, orientation, etc.) will probably change a bit as the setup is moved around, but I've seen integrators out there take on similar projects. So it sounds feasible to me. If you'd like, drop me a line and I can help point you in the right direction. Hope that helps! - Matt

Answered by:
Matthew Breit
Manager - Applications Engineering
Basler AG
Matt [dot] Breit [at] baslerweb [dot] com
(610)-717-5506

Bernard, I'd agree with Matt, so we have developed some Neural vision based software that can learn like a child does, so in theory it would be able to learn good versus bad, the challenge would be how to teach it the different scenerios. Could you send us some images of good and bad, that we could run through the system to see what kinds of results we would get for you?

Answered by:
Andy Long
CEO
Cyth Systems, Inc.
andy [dot] long [at] cyth [dot] com
888-508-7355

Hi Bernard; Just an adder to Matthew's earlier comment. Yes, possibly feasible on the surface. However, even if the inspection is successful covering the difficult imaging and part variations the key question is: Does the geometric relationship between mating surfaces actually confirm a good connection? Machine vision will only (at best) tell you geometrically how "close" one surface is to another. What does one do about variations in the mating surfaces, and what accuracy would one need to ensure mating. This consideration is similar to that faced when inspecting solder or weld joints - the presence and/or amount of material can be gauged, but it does not always indicate a successful joining. I would advise considering a sub-set of this inspection, and would be happy to answer any questions you have if you'd like to get in touch. David

Answered by:
David Dechow
President
Aptúra Machine Vision Solutions
ddechow [at] aptura [dot] com
(517) 655-7860, x11

Have 5 axis, 5'x10' table, CNC router trimming plastic parts. Would like to explore mounting a sensors, cameras, or lasers on the mast that can measure the router bit location to an accuracy of +/- .002. Would also like to put targets or sensors at the extents of the table to measure table location at rest or at a programmed fixed position.

Hi Ken, Just a tip, this would be a perfect application to talk to an integrator about. The company I work for just makes cameras, but I've seen many applications similar to this over the years. I'm not certain if you would need actual cameras, or if you could get away with some simple sensors, but those guys are definitely the ones who can solve the "it depends" questions :) Check the 'Find a System Integrator' link http://www.visiononline.org/search-integrator.cfm. The names I'm seeing there are very reputable and I've personally worked with many of them before. If you get stuck, feel free to email me and I can try to help get you in touch with the right people. Good Luck!

Answered by:
Matthew Breit
Manager - Applications Engineering, Basler USA
Basler AG
Matt [dot] Breit [at] baslerweb [dot] com
(610) 280-0171

Ken, give me a call to discuss your needs, and we'll see if we can help. Alternatively go to the vision show in Chicago next week, and you will be able to see us and all of the manufacturers at the same location. Bring along your samples and get to ask questions of everyone. Andy

Answered by:
Andy Long
CEO
Cyth Systems, Inc.
andy [dot] long [at] cyth [dot] com
888 508 7355

What is the distinction between a Vision System Dealer and a Vision System Integrator

There are companies in the machine vision space where both of these labels could be applied interchangeably. However, in general, a Vision System Dealer (Distributor) sells machine vision hardware. A good dealer (distributor) will have some application engineering resources that can assist with selecting the correct system components to solve the problem at hand and would be able to offer some assistance with trouble shooting after installation. A Vision System Integrator, in general, will be able to offer a much wider array of engineering services. They will be able to take the machine vision system from concept to commissioning and provide on-going trouble shooting and maintenance services. They also, in general, will be able to provide the equipment and services necessary for mechanical, electrical and software integration of the machine vision solution into the rest of the manufacturing process.

Answered by:
Scott Gibbs
CEO
Mercator Robotec Inc
scott [at] mercatorrobotec [dot] com
(519) 884-7880

I manage an engineering group at Eastman Kodak company responsible for our line of high speed scanner equipment. We are looking at the potential for certain aspects of our technology to be a good fit for opportunities within the machine vision industry. High speed capture and image processing are key elements of our scanner industry as well as low cost manufacturing. We are looking to arrange a discussion with industry experts to determine what may make sense in this area. We feel we may be able to bring the industry potentially new capabilities and would welcome further discussion.

Dear Robert; Please give me a call or stop into our office (we are in Fairport, NY). I would be happy to talk with you. Thanks Kerry

Answered by:
Kerry Van Iseghem
Founder
Imaging Solutions Group
kerry [at] isgchips [dot] com
(585) 388-5220, x101

Hi Robert - I think I can help. Please call or email. Thanks.

Answered by:
David Wyatt
CEO
Automation Doctor Incorporated
david [dot] wyatt [at] automation-doctor [dot] com
(574) 344-7936

Hi Robert, I would be very interested in hearing more of your developments. Please email me with some summaries or feel free to call. I've been involved with machine vision for over 2 decades. Thanks, Rex Lee, Ph.D. Dr.Lee@pyramidimaging.com

Answered by:
Rex Lee
CEO/President
Pyramid Imaging, LLC
rlee [at] pyramidimaging [dot] com
(813) 984-0125

Hello Robert, I would be happy to discuss opportunities with you. Please contact me at john.phillips@pleora.com at your convenience.

Answered by:
John Phillips
Sr. Product Manager
Pleora Technologies
john [dot] phillips [at] pleora [dot] com
(613) 270-0625

Robert, we had some customers in the area of large scanners (A1 and A0), which where looking for line scanning sensors, but they shifted the projects due to no good sensors around. Could I have some preliminary datasheets of the sensors you are talking of? Regards, Werner

Answered by:
Werner Feith
CEO
Sensor To Image GmbH
wf [at] sensor-to-image [dot] de
49 8861 2369 0

Hi Robert, Universal Robotics is a software engineering company that has expertise in multi-dimensional sensing and processing, motion control, and artificial intelligence. We specialize in 3D solutions based on a software platform for intelligent flexible automation, providing a wide range of solutions for materials handling, logistics and industrial markets. Universal’s products group by elements of sensory-motor interaction: Intelligence (Neocortex), Sensing (Spatial Vision), and Motor Control (Autonomy). Products integrate into complete applications or function independently, and are camera and sensor agnostic. Because of these unique capabilities, we might be able to form a partnership that would position Kodak as the "go to" company for real-time 3D sensing, whether cameras, lasers or structured light sensors or other technologies. Please check us out at www.universalrobotics.com. I look forward to chatting with you. Hob Wubbena 970 223 2844

Answered by:
Hob Wubbena
VP Marketing
Universal Robotics
hobwubbena [at] universalrobotics [dot] com
(970) 223-2844

We are looking for a vision solution to detect possible defects appeared on the injection molded plastic glass used for vehicle instrument cluster panel. We are currently using human visual inspection method but they have difficulty detecting those subtle defects with their eyes so I think there is a way to help them to detect those defects better.

We have developed a system that could work well for you. This was originally designed to work for flat panel manufacturers. It assisted assembly workers in seeing particles and defects on the cover glass as well as the LCD/LED panel prior to bonding the two together. The product is called GlassSpection and you can see a brief description on our web site www.pyramidimaging.com Please contact me for further details.

Answered by:
Rex Lee
CEO/President
Pyramid Imaging, LLC
rlee [at] pyramidimaging [dot] com
(813) 984-0125

Please feel free to contact us, we'll run some samples through our vision lab so you can see the kinds of results you will get prior to engagement.

Answered by:
Andy Long
CEO
Cyth Systems, Inc.
andy [dot] long [at] cyth [dot] com
(858) 342-6795 cell

Dear Sunho - We have extensive experience in defect detection in flat panel glass sheets like the one on your computer screen, the Iphone cover and many LCD / LED HDTV covers. We are in Tennessee often and also go to Asia. Please call to discuss.

Answered by:
David Wyatt
CEO
Automation Doctor Incorporated
david [dot] wyatt [at] automation-doctor [dot] com
(574) 344-7936

You may want to try highlighting the defects by bringing light in on the edge of the part. This is similar to dark field illumination which is used for inspecting glass and clear plastic parts.

Answered by:
Gene Nolte
Custom Applications Manager
Wilco Imaging, Inc.
gnolte [at] wilcoimaging [dot] com
(619) 977-3797

We have an engineering office in Nashville and would be happy to evaluate product samples to determine the viability of the vision application. Can you prepare a set of Good, Bad, and Marginal parts. Marginal parts being the most critical. Please contact us at sales@integro-tech.com if you would like to proceed with a free vision evaluation.

Answered by:
Shawn Campion
Vice President
Integro Technologies
scampion [at] integro-tech [dot] com
(704) 636-9666

Hi Lee, Universal Robotics is a software engineering company that has expertise in multi-dimensional sensing and processing, motion control, and artificial intelligence. Our products group by elements of sensory-motor interaction: Intelligence (Neocortex), Sensing (Spatial Vision, and Motor Control (Autonomy). Since we are camera and 3D sensor agnostic, we have been able to innovate across 3D inspection spectrum from innovates from low-cost 3D using Microsoft Kinects to high-resolution 3D with high resolution structured light sensors, cameras, and lasers. Our Spatial Vision Inspection product provides very unique capability and is scalable in accuracy and metrical capability based on your need. Please take a look at our products at www.universalrobotics.com, and I look forward to talking with you. Hob 970 223 2844.

Answered by:
Hob Wubbena
VP Marketing
Universal Robotics
hobwubbena [at] universalrobotics [dot] com
(970) 223-2844

We are in a new project vision. The new textile printing technology equipment sublimation transfer paper, required for viewing controls to ensure print quality and control of printed colors. It requires a multi-colored cameras that once you enter the original image, to memorize so that every time you are told, capture the printed image and compare the color tones of the two images (original / captured .) If not applicable, it will warn the controller of the printing machine, to proceed to block it.

It will be important to quantify the allowable color tone change and then make sure you can detect it reliably at high speeds. I would first look at 3 CCD cameras to get a true RGB color separation. If using other CCD or CMOS color interpolation technologies, be sure to remember the impact on resolution and the color skew effects of pixel smear. Also consider automatically converting to hue which is less impacted by change in lighting. JAI makes a 3 CCD line scan. You can find others too I am sure. Good luck! I would love to hear how it turns out for my new book, Realities in Machine Vision

Answered by:
David Wyatt
CEO
Automation Doctor Incorporated
david [dot] wyatt [at] automation-doctor [dot] com
(574) 344-7936

I have a need for a vision system that will be able to count steel sheets in a stack. Example - I have a 2.5" stack of sheets 4' x 8' and each sheet is 18GA thick. How many sheets do I have? The thickness tolerance varies depending on the mill the material was purchased from, along with sheet flattness. Using simple math of measuring overall stack height and dividing by thickness of one sheet does not work. Ideal solution would be a vision system capable of detecting the edge of each sheet and adding them up. Thanks in advance.

Roman: At first glance, this strikes me as a difficult application to get consistent results with machine vision. I would like to understand a bit more about the range of part sizes, thicknesses and material types that the inspection system would need to be able to handle. Also, how do you anticipate handling the parts while they are being inspected? If you would like to give me a call, we can discuss the details of your application and some possible solutions.

Answered by:
Scott Gibbs
CEO
Mercator Robotec Inc
scott [at] mercatorrobotec [dot] com
(519) 884-7880

Roman, One first might consider this to be a very straightforward and easily-implemented machine vision application. I have done this inspection before, and would caution you to be aware of issues that may impact the reliability of the counting. With 4'x8' sheets of steel the presentation likely will be a big issue. It might be difficult to ensure that the edges are somewhat closely aligned on the stack so they can be illuminated and imaged, and difficult to present each stack to the camera at a repeatable distance from the lens. Also, changes in the edge cut, and even rust or other debris on the edges can cause confusion in detecting the small spaces between sheets for counting. In general, you must be sure to do sufficient up-front analysis of all of the variability of the stacks in order to arrive at a successful system configuration. I would be very interested in helping your company with this application, and invite you to contact me at your convenience. Good Luck, David

Answered by:
David Dechow
President
Aptúra Machine Vision Solutions
ddechow [at] aptura [dot] com
(517) 655-7860, x11

Roman, Have you considered using a 3D vision system to monitor the gaps between each sheet, this might be a good fit to solve the different problems you are going to have with a 2D machine vision solution. If you're interested we can talk further to see if this would be an option. One other question, will you be doing this counting outdoors? Andy

Answered by:
Andrew Long
CEO
Cyth Systems, Inc.
andy [dot] long [at] cyth [dot] com
888 508 7355

Dab side of the stack with a 2" wide dye to enhance the edges. Setup your camera to be sensitive to the dye when viewed at a 90 degree angle. Use programmed lighting to take two shots of the stack under different lighting conditions. Use software to conclude what the sheet count must be based on the detected edge pitch, the total height and the number of sheet edges seen. Call me and Ill be happy to quote you a hand held unit.

Answered by:
David Wyatt
Certified Vision Professional - Advanced Level
Automation Doctor Incorporated
david [dot] wyatt [at] automation-doctor [dot] com
(574) 344-7936

I would like to specify the required accuracy of a vision system measurement. I historically have used the average of the absolute values of the errors being less than 1/20 of the total tolernace as a guide. Is this a good indication fo the accuracy of a vision system measurement? I also use six times the standard deviation of a repeated measurement to be less than 1/10 of the total tolerance as a guide for capability. Is this a good measure as well? If not what are some good ways to verify accuracy and capability based on using measured data compared to tolerances and know measured values?

First, I want to clarify how I read your question with respect to the term "accuracy", and hopefully I'm on the right track. Accuracy is how close an observed measurement is to the "actual" or "true" reference value. Evaluating the "accuracy" often is less important (or achievable) than evaluating the "precision" or "degree of reproducibility". The second calculation that you mention is one measure of precision. Another very widely used measure is a Precision to Tolerance or P/T ratio. The calculation is: six times the standard deviation of a set of repeated measurements on the same part, divided by the tolerance (6 * measurement sample sigma / usl-lsl). Multiply by 100 to get a percentage. A result of less than 10% is a capable system, but 10-30% could be acceptable depending upon the application. This is just one other option. In general I'd point out that there are no absolutes here, and some judgement call is necessary based upon the application. You're on the right track.

Answered by:
David Dechow
President
Aptúra Machine Vision Solutions
ddechow [at] aptura [dot] com
517-655-7860 x11

Jason -- There's a lot to say to answer your question. My response will be very brief. There are two parameters: accuracy and repeatability (aka resolution). Accuracy -- the average difference between the "true" dimension and the measured dimension. Repeatability -- the standard deviation of repeated measurements on one part. In classic QA metrology accuracy must be 1/10th the tolerance and repeatability must be 1/10th the accuracy. However, this protocol is violated with regularity. In machine vision, because we have eliminated the human factor that was considered in setting the above ratios, we typically use smaller ratios. My preferences are to use 1/5th to 1/4th as reasonable for a final ratio of repeatability to tolerance of around 1/20th. Hope this helps. Contact me if you have more questions.

Answered by:
Perry West
President
Automated Vision Systems
perry [at] autovis [dot] com
(408) 267-1746

David Dechow and Perry West do a great job in explaining the accuracy and repeatability issues surrounding vision. Bear in mind that you need to know the maximum variability and mean shift in your overall vision results over time. Unfortunately, you will not know this until you implement the system. Many people simply state that the system must pass a Measurement System Analysis (MSA, a specially designed experiment that seeks to identify the components of variation in the measurement). It can take a lot of money to prove or disprove that any system, which includes part variability and presentation variability passes or fails an MSA. Then you may find that the vision system is not a major factor. Also, remember that you will need a calibrated tool to compare the vision system to, such as another optical comparator or CMM. We do this on a contract basis everyday. Call us if you need any help.

Answered by:
David Wyatt
Certified Vision Professional - Advanced Level
Automation Doctor Incorporated
david [dot] wyatt [at] automation-doctor [dot] com
(574) 344-7936

We have operators standing by our metal stamping machines and monitoring that nothing appears to have changed in the process. Can we get a vision system to replace this duty without too much complication?

Hi John, The key will be to quantify what "nothing appears to have changed" means in numeric terms. A vision system is a calculator with a camera attached to it. We will need something to calculate. Size, shape, color, reflectance etc. To be successful we need a complete list of what is expected to be monitored and what the max and min numeric limits are for each item on the list. I can walk you through this process for free it you call me direct. It will be somewhat complicated, but the complicated part will be defining what to monitor and how much it can change. The vision side is easy in comparison. Let me know if we can help,

Answered by:
David Wyatt
CEO
Automation Doctor Incorporated
david [dot] wyatt [at] automation-doctor [dot] com
(574) 344-7936

Hi John, I agree with Mr. Wyatt. If you could provide us with more details we could assess you on whether laser triangulation is a good option for your project or else. Best regards,

Answered by:
David Sabaté
Marketing
AQSENSE S.L.
dsabate [at] aqsense [dot] com
34 972 183 215

Hi John, There are a number of application approaches that can be applied to solve the described stamping inspection application. The statements below assume thru hole and perimeter stamping. The most aggressive method is to reproduce your stamped part 2D drawing within the vision system. This method permits your QA department to directly correlate the expected versus actual results geometrically. The second method is to apply a print quality inspection toolset to highlight minor difference between the golden template (master) and the produced parts. A print quality inspection system can be trained with a single master part or a series of masters to provide a statistical model for grading. If a topography map of the part is required this is achievable but attention must be paid to the material finish and how this may affect the laser systems.

Answered by:
Shawn Campion
Vice President
Integro Technologies
scampion [at] integro-tech [dot] com
(704) 636-9666

This depends on the area of overall inspection. If there is a clear shot to the product, it could be very simple to mount a camera box above it to inspect. We build customized software with our machines so its very adaptable. What are dimensions of the area you are searching? If you can provide me with some more details, we may be able to assist you. We are local (Glen Ellyn, IL) and that being so, could come look at exactly what you are inspecting if you are looking towards purchasing a solution.

Answered by:
Jeff Griswold
Engineering Technician
Silgan Vision
jeff [at] mcsvision [dot] com
(630) 790-4497

I'm working for solar cell company. We are looking for a tool for macro vision inspection. I wonder if, although the color of processed wafer is different between light green to dark purple, you can check wafer crack (around 3~10mm) or not. Because our process is not really fix yet and I suspect that the wafer color will border to use vision system. But I want to use vision crack inspection tool. Therefore please recommend which tool is good. Thanks.

We have been able to do this with wafers of various materials. If you can send us a sample then we can demonstrate our capabilities to you. Please contact me for more information.

Answered by:
Rex Lee
CEO/President
Pyramid Imaging, LLC
rlee [at] pyramidimaging [dot] com
(813) 984-0125

Hi All, I am looking for some ideas on a 3D vision system that will profile and create an image of food stuffs travelling along a conveyor and the data of the profile to be then sent onto another system possibly PLC which will then, based on the data, decide on how best to portion this up. The product may be up to 800*200*150mm and up to 30 per min. Cost of a system is highly relevant and the system needs to be able to deal with uneven product surface. Thanks.

We are a machine vision integrator with Offices in California and a Support Office in Oxford UK. We deal with a large number of 3D systems, I would like to discuss your project in more detail. I am actually based in California, so 8 hours behind you, but I could give you a call early my time. We usually put products through our vision lab first to ensure that we are able to give you the data you need, and show you a report prior to undertaking a contract. If you're interested let me know. Also the Vision show is coming up in Boston in early May, if you're thinking of investing in this kind of technology it may be a good show for you to attend. We will be exhibiting there, so would be willing to host you and introduce you to the manufacturers, see the events section on Vision Online for further details.

Answered by:
Andrew Long
CEO
Cyth Systems, Inc.
andy [dot] long [at] cyth [dot] com
+1 (858) 537 6541

Dear Robert, Our company, AQSENSE, have built this exact kind of machine for leading food companies using our 3D machine vision library SAL3D. You can contact me so that we can elaborate on the requirements and costs of the product.

Answered by:
David Sabaté
Marketing
AQSENSE S.L.
dsabate [at] aqsense [dot] com
34 972 183 215

Hello Robert, There are a couple approaches to take on this type of application, depending on the accuracy required. The most cost effective is a 3D CMOS approach, this is similar to the technology used in newer elevators, higher end autos for smart air bags deployment, and smart automated door sensors. These systems use time of flight scanned laser imaging to produce a 3D image. SICK and IFM Effector have industrial sensors that use this technology, I'm not sure if they have the resolution you need. We have access to this technology and can provide you a turnkey solution. Hope this helps.

Answered by:
Keith Anderson
Software Engineering Manager
Integrated Packaging Machinery, LLC
keith [at] callipm [dot] com
(616) 502-4935

Hello, My company is looking for a way to verify the 2D dimensions of a flat part against a standard shape. This would be for a semi automated production packaging environment, so I am not looking for any human adjustment during the part validation. I need help identifying vendors of this type of equipment and or someone to bounce ideas off of. Thank you.

Due to the many variables of your application it would be great to have a conference call to discuss the particulars. How many parts, size, and what sort of processing you want to have happen. Please contact me so that we can set up a time to go over the preliminary Q&A session.

Answered by:
William Corns
VP of Sales
Wilco Imaging, Inc.
wcorns [at] wilcoimaging [dot] com
(858) 689-9050, x3202

We have an office in Nashville, TN staffed by an application engineer with 30+ years in the machine vision industry that could review your application and provide hardware & software recommendations or we could quote a turnkey solution. Please contact us at sales@integro-tech.com.

Answered by:
Shawn Campion
Vice President
Integro Technologies
scampion [at] integro-tech [dot] com
(704) 636-9666

Please contact me directly if you would like to discuss your project requirements and some solution options.

Answered by:
Scott Gibbs
CEO
Mercator Robotec Inc
scott [at] mercatorrobotec [dot] com
(519) 884-7880

If you send us samples, we typically run them through our vision lab and show you the kind of results you will achieve. We will do this prior to any contracts changing hands, as we only want to be involved in successful projects, and this is a way of ensuring that. Are you looking for a shadow-graph type system where you only care about the silhouette or are you interested in more dimensions? We deal with 3D systems as well as 2D. We are HQ in San Diego, CA. Best of luck, feel free to contact me if you'd like.

Answered by:
Andrew Long
CEO
Cyth Systems, Inc.
andy [dot] long [at] cyth [dot] com
+1 (888) 508 7355

Click here to view archived questions for this category.

Archived Questions

Click a question to reveal the answer. To return to the the latest questions for this category, click here.

We are trying to find a vision system that is eligible for OCR. The piece that is to be read is a serial number (17 characters) which is engraved into a machined metal part . The part will be manually moved, so speed is not a priority. I am not too familiar with vision systems, so I was wondering what would be the best for this type of application?

We have successfully deployed 90+ OCR systems last year using the Cognex In-Sight platform and the new OCRMax algorithm. I believe this a best in class algorithm. Based on the described application, when configured properly the solution should perform very well. Do you have any samples images or parts that you can sent to us for review and comment?

Answered by:
Shawn Campion from Integro Technologies - Vice President
scampion [at] integro-tech [dot] com
(704) 636-9666

Jodie, OCR tends to be tricky to get 100% success with. It is usually easy to show a demo, but repeatability is usually the issue you need to look out for. May I suggest you send samples out to either manufacturers of equipment or to Systems Integrators like ourselves. This will allow us, or companies like us to evaluate your samples, and see if we are likely to have a reliable system. We will run the samples through our vision lab and give you a report all without commitment from either side. This allows both parties to know what they are getting into. We want to know its feasible, and you want to know its feasible. I appreciate that this means you have to do more work in getting the samples together, but trust me, more work now, less stress later. Good Luck.

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
888 508 7355

Hi Jodie, We offer a cost effective, complete vision system in a camera, from Baumer called VeriSens. It is designed to read any number sequence without having to train on the font. I invite you to contact us. There is no substitute for a live conversation when it comes to explaining challenges and discussing solutions. We can talk about lighting an optics as well. These are two of the most important considerations for any vision system.

Answered by:
Gene Nolte from Wilco Imaging, Inc. - Custom Applications Manager
gnolte [at] wilcoimaging [dot] com
(619) 977-3797

We are nearby Shanghai, China; we want to set up a machine vision checking station for porosity (diameter 0.4~1.0 mm) checking after machining of Al part; and these pores are located in the cylinder hole of part; please contact me if you have a sub-company or agency which is in China, so we can take a further communication. Thank you.

LUSTER LightVision Tech is one of Pleora's representatives in China. Please contact Simon Wu at xinranwu@lusterlighttech.com.

Answered by:
John Phillips from Pleora Technologies - Sr. Product Manager
john [dot] phillips [at] pleora.com
(613) 270-0625

I am working on a process for inspecting a 52" wide white web of fibers for any type of non-white contaminants. Ideally we'd like to detect any non-white item in the web that is 700 microns or larger. Due to upstream parts of the process, the line speed varies but is never greater than 20 yards/minute. The web is backlit and appears and is translucent in appearance (100 mils thick). Given the relatively slow speed that we are running, is it possible that we might employ used equipment (from 4-5 years ago) to reduce cost and still have an effective system. If so are there any integrators who would even be willing to work with older equipment. Any feedback is appreciated.

You are better off with new hardware for life cycle. What kind of target budget you have? Since you are looking for a small defect across a large web, the cost is dictated by resolution requirements. We will be glad to work with you to find the cost effective solution. If you are interested, please email your contact information so that i can call you. thank you

Answered by:
Reddiar Anbalagan from Automated Vision, LLC - President
anbu [at] automatedvision [dot] com
(608) 836-9393

There are two parts to system cost -- parts and labor. Using used equipment might save a bit on cost, but nothing on labor. Often, labor is the larger of the two costs. Perhaps a bigger question is what economic life do you need from the system. Components become obsolete and no longer available for maintenance. Using older components shortens the expected economic life of a system. If you design with current components, you can reasonably expect 7 to 20 years of economic life. After that, the system would need to be redesigned with current hardware if the function is still required. If you use, equipment that is 5 years old, you sacrifice much of the economic life of the vision system. Considering the small cost savings and the rather larger loss in economic life, it rarely pays to design a machine vision system with old components.

Answered by:
Perry West from Automated Vision Systems - President
perry [at] autovis [dot] com
(408) 267-1746

We would be willing to work with older equipment. However, I would echo the comments from the other responders that when the total life cycle cost of the system is considered, new hardware would probably be more cost effective. Please contact me directly if you would like to discuss some options.

Answered by:
Scott Gibbs from Mercator Robotec Inc - CEO
scott [at] mercatorrobotec [dot] com
(519) 884-7880

I would like to echo the previous statements made. The cost of labor to undertake an inspection usually far outweighs the hardware costs. One of the things I would suggest you do is provide some samples of good and bad, and the better you are able to classify your defects the more chance of producing a reasonable spec. Reddiar hit the nail on the head, when he said your system cost is dictated by resolution, and very often our customers do not think in resolution, they think in terms of defects; that I need to detect, and those that I would live with. If you can get samples of the defects you can live with, and those that are too small to worry about we would gladly run them through our vision lab, and give you a report that will help you decide. Let me also add here that you've got through the first hurdle, your asking questions in the correct forum, which ensures that you should get the knowledge that the people answering these questions can provide. Good Luck.

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
888 508 7355

Hi LC, I attempted to contact you the other day after you left a message at our NC office. I will try again today. Please forward your email address.

Answered by:
Shawn Campion from Integro Technologies - Vice President
scampion [at] integro-tech [dot] com
(704) 636-9666

Dear Sirs, We, are investigating an embedded vision systems, rugged and wireless, allowing realtime image per second display and zoom, on 10" monitor. Our actual research leads to wired systems, and we would like to know if some of your Members will eventually propose wireless cameras offering real time imaging, and autonomous power supply (battery, solar, wind ...).

Dear Sebastien, We (MoviMED - Custom Imaging Solutions) specialize in designing custom vision systems for situations where you can not find a suitable off-the-shelf solution. We have done some wireless vision solutions before for autonomous robotic vehicles, surveillance systems as well as for unmanned aerial vehicles. I would love to discuss your application further. Please contact me directly.

Answered by:
Markus Tarin from moviMED - President and CEO
m [dot] tarin [at] movimed.com
(949) 699-6600

Sebastien, The challenge is obviously not wireless transfer of data, we have all been doing this for quite some time, the challenge is real-time image transfer, and what is the consequences if the data is not sent in time. We have systems that transfer image data wirelessly, but they are only used, when are user requests a transfer, or where wireless is the only option, and if the data doesn't arrive, it doesn't really matter. The other question to consider is the data transfer protocol, UDP or TCP. TCP is what most networks are based on, and it guarantees delivery, which means it takes longer to transfer data, and you can see significant lag. With UDP the data is sent out, and if it never arrives, nobody ever knows, but with images if there is another right behind it, then maybe you don't care. How this starts to address your question. I'd be interested on what some of the camera manufacturers have to say about including wireless in the cameras in future.

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
+1 (888) 508 7355

Hi, I am trying to get a quick answer about which type of light (RGB LED, IR or else) is suitable to the pattern recognition of the following components in my assembly machine: 1. The edges of a transparent glass. 2. Gold plated metal surface. 3. Etched silicon surface (there is some rainbow colors if use blue color light). Highly appreciate any prompt help! Wei

Unfortunately, you have fallen into the common trap of considering only the object being viewed when selecting the illumination. You need to consider contrast -- what you are trying to discriminate. For example, you are looking at a gold plated surface, but what are you trying to image? Looking for missing plating may take a different lighting scheme than looking for surface irregularities. I'm in the bay area. Why don't you call me and we can discuss this further.

Answered by:
Perry West from Automated Vision Systems - President
perry [at] autovis [dot] com
(408) 267-1746

Hi, I think that your question is a bit too general. There are numerous ways to highlight surfaces for pattern recognition. It is also best if you control the entire environment so that you can stop stray light. It would also be best to match the lights used with the camera's sensor and that sensor's wavelength sensitivities. The general goal is to make what you wish to see the part of the image that is obvious while diminishing all other parts of the image. So, if for 1 it is best to make everything look black except for the edge. If you fully control the environment then you could do this with any structured light and orienting the camera carefully so that light from the edge goes to the camera and nothing else appears. For 2 is there a pattern in the gold metal surface that you wish to see? Is it etched into the surface or raised above it. Either way a dark field light would be best. Same for 3. above if all you wish to see is the pattern that is etched into the silicon.

Answered by:
Rex Lee from Pyramid Imaging, LLC - CEO/President
rlee [at] pyramidimaging [dot] com
(813) 984-0125

We are looking for the vision systems for following missions. Do you have a capabilities and references for meeting following tasks? Quality management system which can determining surface defects on automatic transmission case for passenger cars at mass production maching line. Defects could be chip, burr, porosity, dent, oil contamination, scratch, crack, misaligned hole and hole opening.

Chong, we have made a number of systems for inspecting surface defects. Could you send me an email or a phone call to discuss your particular needs. What we like to do is get samples from you that we can run through our lab, to show you the kind of results you will get. Andy

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
+1(858) 342-6795 cell

This sounds like a good application for a PPT Vision system. Please contact our partner in South Korea, Vistec +82 226302580 vistec@vistec.co.kr They should be able to help you your application. If you have any questions, please let us know.

Answered by:
Steve Maves from Datalogic Automation - Manager of Application Engineering
stevem [at] pptvision [dot] com
(952) 996-9504

Dear Mr Chong, Our standard library SAL3D for 3D acquisition and manipulation can be a perfect tool for your application. It is widely used for quality inspection applications and, for example, the leading German vision magazine Inspect published an article regarding the topic (Original German article in INSPECT 3/2011): www.aqsense.com/docs/surface_inspection_inspect.pdf I'm sure that Cyth Systems could use SAL3D for the tests in their lab. Best regards,

Answered by:
David Sabaté from AQSENSE S.L. - Marketing
dsabate [at] aqsense [dot] com
34 972 183 215

Do you have a formal specification for this project?

Answered by:
Shawn Campion from Integro Technologies - Vice President
scampion [at] integro-tech [dot] com
(704) 636-9666

Hi, I have certain questions to be answered for my study on vision inspection in food and beverage industry. Can anyone help me with the size of the vision inspection market in US and Europe and the growth rate over around five years.

Anurag, The AIA has studys that describe these markets, (you may have to purchase or join AIA). Andy

Answered by:
Andrew Long from Cyth Systems, Inc. - Senior Director
andy [dot] long [at] cyth.com
+1 (888) 508 7355

I am looking for a camera or video system that can be displayed to a Monitor. I would like to have a bulls eye or some sort of target system to be displayed on the monitor that can zoom in or out to focus. The purpose is to put a small pin through a hole. Thank you.

Hi Bill, If you are thinking about using a GigE Vision compliant camera, then you might be able to make use of our vDisplay HDI-Pro (http://www.pleora.com/our-products/vdisplay-video-receivers/vdisplay-hdi-pro). It takes video off the gigabit Ethernet network, and displays it on an HDMI or DVI-compliant monitor. As for adding the target, you can do so by routing the image from the camera through a PC using our eBUS SDK. Receive the image on the PC, apply the target, and then resend the image on the network to the vDisplay HDI-Pro. Because it operates on Ethernet, the distances between the camera, PC, and vDisplay can be quite long (up to 100m, longer if using a switch). The SDK works on both Windows and Linux operating systems. Alternatively, we worked with a company in the past who had purchased lenses with a target embossed directly on the lens. Unfortunately, I think that company (the lens manufacturer) is out of business, but there may be others.

Answered by:
John Phillips from Pleora Technologies - Sr. Product Manager
john [dot] phillips [at] pleora.com
(613) 270-0625

Bill, I think it depends on to what extent you want a DIY project. If you buy a camera from a vendor with a basic SDK they'll likely have video viewer sample code. I know Basler and the Imaging Source do. To implement zooming and draw the target I'd do it in software. You can stitch in an external library to your application and leverage that code. I'd suggest OpenCV or Emgu CV (an OpenCV wrapper) to do that heavy lifting. You could build it yourself or you could hire us (or someone else) to do it. For basic zooming and a target, it'd take a software guy about 5 days to put together in C++ or C#. You'd likely want to modify and rebuild OpenCV or Emgu CV to always put your target in the middle of your zoomed image. You could also buy something more off the shelf like John suggests below and set that up. I really don't know if the exact product you want exists though. If it's a fixed station you need something by Scienscope could work.

Answered by:
Tom Brennan from Artemis Vision
tbrennan [at] artemisvision [dot] com
(303) 832-1111

Hi Bill, People often use our Vision Systems for these kinds of semi-automated vision tasks. Please feel free to drop us a line and we'll connect you with one of our partners. Thanks, Steve Maves stevem@pptvision.com 952 996 9504

Answered by:
Steve Maves from Datalogic Automation - Manager of Application Engineering
stevem [at] pptvision [dot] com
(952) 996-9504

What products should we use to clean the sapphire glass on our IM-6000 inspection system?

Hi Kenneth - I believe the IM-6000 is a Keyence product. They have great tech support. I would call them at 888-539-3623 to be sure you get the factory approved answers to your question. Kind Regards,

Answered by:
David Wyatt - Certified Vision Professional, Advanced Level from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

I have to agree with David. Ask the manufacturer. If it were only sapphire, then any cleaner would work. It depends, though, on whether there is a coating on the sapphire (e.g., anti-reflective coating) and what other materials are around the window. The first step, regardless of materials, is a gentle blow-off with very clean air -- canned air or a squeeze bulb are common sources (not most factory compressed air). Generally, an reagent grade acetone wipe followed by a cleaning wipe with reagent grade isopropyl alcohol. The right technique should be used to avoid scratching the optical surface with grit removed from the surface. There are other preparations sold by companies such as Edmund Optics that might be more suitable.

Answered by:
Perry West from Automated Vision Systems - President
perry [at] autovis [dot] com
(408) 267-1746

I am investigating the use of a vision system to quantify printing defects produced by thermal transfer printing on polymer labels.

Hello Hugh - I have done this several times. First identify your minimum defect size and colors. Next identify how much distortion you have from label to label. Finally decide how many label types you want to run and how much effort you are willing to put into the setup each label. Select a system that uses a pattern matching technology and most importantly get the vendor to prove it out in advance for free (or a small fee) on stored images. Feel free to photoshop the stored images to challenge the prove out.. We can do this for you at our standard rates if you do not have time. Good Luck and call to discuss if you wish.

Answered by:
David Wyatt - Certified Vision Professional, Advanced Level from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

We would be more than glad to assist you with your machine vision application please contact Integro Technologies at 704-636-9666 or via email at sales@integro-tech.com

Answered by:
Shawn Campion from Integro Technologies - Vice President
scampion [at] integro-tech [dot] com
(704) 636-9666

Hello Hugh - There are a large number of vision vendors who consider this a "bread-and-butter" application, including my own. I am more on the technology side, but you will find that different vendors begin to sort out when having to read sub-optimal printing, slanted letters, dot-matrix and the like. The basic application can range from easy to impossible depending on the print quality. Is this OEM? Later, Tim White

Answered by:
Tim White from Mettler Toledo CI-Vision - Vice President, Technology
tim [dot] white [at] mt.com
(630) 446-7761

Hugh, The first thing you need is a good set of samples and a bad set of samples. We are a machine vision integrator, and we find that if the customer is able to get these comprehensive sets, then the project usually develops more easily. If you have these samples, we can run them through our vision lab, and see how complex your problem is. Andy

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
+1 (888) 508 7355

Does a benchmark study about delivery standard times for developing, in general, a vision system exist?

Hi Jesus - I am not aware of a study that benchmarks the delivery time of a standard vision system. My employees, customers and investors are always asking this question. The problem is that there is no standard vision system. A vision system is quite often an effort to select hardware and write software for images that are not fully understood in high volumes. As a result it often takes several iterations to develop a successful solution. In the construction industry it is well understood how long it takes to lay a brick wall of a certain length, width and height. The development of a vision system is filled with unknowns. All of that being said, simple systems can be done in a week.. Medium systems may take several weeks. Difficult systems takes months. Complex systems can take a year. Some systems are never really finished. Every image presented for analysis is unique. You bring up a great question. I hope this helps

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

What research can we conduct using vision systems (machine vision) in agriculture today?

When you say agriculture, I understand this to be in the field rather than food sorting done in a plant. There are classic areas of application: automatic harvesting, field sorting (e.g., removing dirt clumps and foreign material), weeding and selective application of herbicide, and even pruning (of grape vines). Any of these tasks could be extended to new field crops that have not been adressed before. There is the whole emerging area of smart farming. Evaluating the growth and health of plants to make a map of the field for local optimization of fertilizer, pesticide, herbicide, and irrigation application. Autonomous navigation of field machinery may be one area where progress can be made. I do know that some effort has been made in that area.

Answered by:
Perry West from Automated Vision Systems - President
perry [at] autovis [dot] com
(408) 267-1746

I am looking for largest machine vision integrator in the world for industrial manufacturing. By largest I am hoping to find largest in terms of the number of cameras deployed in manufacturing applications. I am not looking for the neither the largest user, nor the largest camera or software manufacturer. Any Idea?

I do not know who would qualify any claims by vision integrators on a global basis as being the largest based on the number of deployed systems. There are a number of well qualified and experienced machine vision integrators in North America who could compete for this title. Integro Technologies has been in the top 3 integrators for Cognex Corporation the past three years and we are in position to be 1st at the close of business 2011. This year we have deployed +300 vision systems across a wide range of industries. I would have to research our records to get an exact count. We continue to expand our 2D and 3D solution portfolio and hardware platforms.

Answered by:
Shawn Campion from Integro Technologies - Vice President
scampion [at] integro-tech [dot] com
(704) 636-9666

Hello Stephen - Since you will probably need more than one quote, look up Automation Tooling Systems, http://www.atsautomation.com/automation/assembly/as_vision.asp I am a past GM employee from Vision's early days, Specifically, Delco Electronics Equipment Development Vision Group as well as GM Tech Center Machine Intelligence Chairman and Hughes Aircraft / GM Machine Vision Liaison. If I can help from a consulting or project management standpoint, please contact me directly. Thanks, David Wyatt Certified Vision Professional Advanced Level

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

Hello Stephen - I got started in one of the original GM-funded vision companies, AISI in Ann Arbor, in 1981, and have worked for a variety of different mfg's, industrial end usesrs and integrators since then in various capacities. The integration company I work for was acquired by Mettler-Toledo a couple of years ago, a $2 billion swiss-based company. Maybe not the biggest integrator ever, but I have seen a lot of cameras put in a lot of places doing a lot of things. Best regards, Tim White

Answered by:
Tim White from Mettler Toledo CI-Vision - Vice President, Technology
tim [dot] white [at] mt.com
(630) 446-7761

I have a requirement to build a vision based gauging system where I need to achieve one micron least count and accuracy. Typically how is this achieved, what kind of camera and controls are used to build a machine?

A typical 6-sigma repeatability for a CCD array camera is 1/3 of a pixel. For example, a typical 5meg camera has an array size of 2500 pixels in X. If the field-of-view is 25mm wide, then each pixel has an X dimension of 0.01mm. The expected 6-sigma repeatability of the vision system, assuming a high-quality edge image, would then be 0.01/3 = 0.0033 = +/- 0.001666mm. This can be improved upon by taking a number of measurements in sequential FOV's, and using the average of such measurements as the "measurement". Accuracy improves as the square root of the number of measurements averaged, so that averaging 25 measurements will nominally yield a 5X improvement in precision. At micron-scale measurements, the measurement accuracy is quickly overwhelmed by part-to-camera distance variance and other factors. Even with a quality telecentric lens, such effects largely determine the obtainable effective measurement precision of the system.

Answered by:
Tim White from Mettler Toledo CI-Vision - Vice President, Technology
tim [dot] white [at] mt.com
(630) 446-7761

Hi, I am very interested in integrating a computer vision system into my warehouse. I have spoken to a few companies and am in the initial stages of trying to see if computer vision can be integrated into my production. Without a doubt I am sure if it has not reach a point where it can be integrated, the technology will get there in a few years. My main goal is to have computer vision technology be able to learn and differentiate between clothes such as t-shirt, pants, jeans, shorts, boxers, underwear, shirts, etc. when traveling down a conveyor. If you know someone who can help me I would love your referral. Also, I travel to third world countries often and am seeking to start a new venture in a foreign market. Thus, I think with Computer Vision can offer great competitive advantages in existing business. I would like to know what is the biggest computer vision conference as I would like to attend and get more information. I look forward to your response and working with computer vision technology in the very near future. Thanks.

Hello, I would like to address the vision conference portion of your question directly. The AIA will host a four day educational conference in conjunction with The Vision Show in Boston, Massachusetts in the US. The dates are May 7-10, 2012. We have a comprehensive program with four full days of training. Our agenda is not yet final, but will be after the beginning of 2012, check back on Machine Vision Online "events" to see the courses and to register. You'll also have an opportunity to meet many of the industry's leading system integrators when you are there. We hope to welcome you in Boston.

Answered by:
Dana Whalls from AIA - Vice President
dwhalls [at] robotics [dot] org
(734) 994-6088

Hello Mr. Khanji, What you want to do is doable with the machine vision technology. Our company designs customized machine vision systems to meet the specific needs of customers and I think we can do it for you. However we need to know more specifications about your desired system. What is the speed of conveyor? Are clothes separated by a space when traveling on a conveyor or they might be stuck together? I will be happy to discuss this application with you if you send me an email or call me.

Answered by:
Kasra Ravanbakhsh from Kasra Hooshmand Engineering Co., P.J.S. (KDI) - Managing Director
kasra [at] kdi-co [dot] com
98 21 8867 8023

So much depends on how the clothes are placed on the conveyor. If they are carefully placed so that a t-shirt always looks almost identical, then machine vision may have a solution. If the clothes are just randomly placed on the conveyor in a lump, then it may be extremely difficult to identify. Even if the first condition is met, there will be considerations of different sizes and different colors. Color may pose the greatest challenge -- there must be contrast between the item and the conveyor. However, the latest generation of software is getting quite good at dealing with low contrast. I would be interested in helping you scope out the requirements for your system and finding if there is a technology that can work reliably for what you need done. Please contact me to discuss this in more detail.

Answered by:
Perry West from Automated Vision Systems - President
perry [at] autovis [dot] com
(408) 267-1746

Hello Mr. Khanji, I will be in Houston Texas the first week of December. I would be happy to meet with you to discuss your options. Please contact me directly. Sincerely, David R. Wyatt CEO Automation Doctor Inc AIA Certified Vision Professional - Advanced Level 574-344-7936

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

Hi Imran, Please contact Integro Technologies via email or phone to discuss your textile machine vision application in greater detail. sales@integro-tech.com 704-636-9666 www.integro-tech.com

Answered by:
Shawn Campion from Integro Technologies - Vice President
scampion [at] integro-tech [dot] com
(704) 636-9666

Hello, I need to measure traffic volume and speed of vehicles. Is this something imaging HW and SW can do? Can you give me a few companies that may supply the necessary cameras and software? You have a large list of companies, so some direction will help.

Dear Eric; Feel free to give me a call to discuss your application. We make cameras for traffic applications for OEM's throughout the world. Thank you, Kerry

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

Try Perceptics, www.perceptics.com. Randall Henderson, EDT Inc.

Answered by:
Randall Henderson from EDT - Senior Software Engineer
randall [at] edt [dot] com
(503) 690-1234

Eric, If you want a system delivered, feel free to contact myself. We can build a system to your specs in terms of what level of detail you want to deal with. Also we use a number of differing technologies, based on environmental conditions etc.. Andy

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
888 508 7355

We have many small labels on our products. As part of Quality Assurance we go through and read all the labels to check for typos and for proper location. The labels are just plain text. What type of system would work best, at the minimal cost, to be able to point a gun/camera at these labels, and have it convert the image it sees into the text that is on the label. The software would need to have an API that is programmable using the .NET Framework from Microsoft.

John: A quick Google search for "handheld OCR" produces numerous devices that will no doubt meet the minimal cost criteria and will easily plug into a PC via a serial or USB port. Many of these devices are not particulary rugged and therefore are not suitable for use on the factory floor. However, they may be suitable for your purposes. Please contact me directly to discuss your application requirements in more detail.

Answered by:
Scott Gibbs from Mercator Robotec Inc - CEO
scott [at] mercatorrobotec [dot] com
(519) 884-7880

John, I noticed your question, and I wanted to pass some information your way. It is very common to focus on the software as the solution to solving machine vision problems, we are a machine vision integrator, so we use lots of software to solve problems. But the key to a good vision system is a good quality image. And I want to describe what that means. You need maximum and consistent contrast between the characters and the background. You need consistent fonts within the image. I would warn you against assuming that because the human eye - brain combination can instantly see it, that does not mean that a machine vision system can. I would ensure that you work with a lighting company and an optics company to make sure the images you get in the system give you a reliable consistent output. We make our engineers prove on the simplest task that the image is "good" by running software at it. I hope this helps you, please don't underestimate the first piece, getting a good image. Andy

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
888 508 7355

John - To add to what the other experts have said, remember that you will get the best results using an optical character reading font such as OCRA. Yes you can train a font, you can have many fonts trained at the same time and even use neural nets to help detect and learn the fonts. My experience has been that those methods have varying degrees of success and it takes a long time to train and retrain and retrain and retrain fonts. The more variance you allow in a font the lower the read rate and the more susceptible the system is to confusion. I recommend using a 2D barcode along with your human readable printing. It looks modern and works the best in a changing environment. David R. Wyatt AIA Certified Vision Professional Advanced Level

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

John, We sell the Microscan 2D Barcode scanners and have used these in multiple installations. We could send one your way for testing with your application. Let me know how I can help.

Answered by:
William Corns from Wilco Imaging, Inc. - VP of Sales
wcorns [at] wilcoimaging [dot] com
858-699-2518

I'm trying to build a Tomato Sorting Machine.What brand & model of vision system can be perfectly used to inspect the size & defect marks on a tomato ( the tomato is moving on the conveyor ) ?

Hello Lee, It is not about the brands, it is about the philosophy of design. You should first consider, based on which characteristics of tomatoes, i.e. color, size, ripeness, spots etc..., you want to sort them. Then you should calculate the requirements for cameras based on the requirements of your system, i.e. resolution, type of camera as color/gray scale/non-visible, sensitivity and speed of sensor among other things. You should also be an expert in software programming for machine vision application, e.g. be familiar with a professional machine vision software library. Anyway if you need to outsource the development of this sorting machine, you can always count on our company!

Answered by:
Kasra Ravanbakhsh from Kasra Hooshmand Engineering Co., P.J.S. (KDI) - Managing Director
kasra [at] kdi-co [dot] com
98 21 8867 8023

Dear Lee Swee Sin; We have worked with a few sorting firms with very similar application as your project. Please contact me via email and I can provide further information.

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

Lee, Have you undertaken any vision system development in the past. I am not aware of a "tomato sorting machine" that is available off the shelf. You may have to invest in software development time, using an adaptive type technology to solve this problem. This would involve a system that is trained to look for defects, by showing it the defects. You may also need to use 3D type technologies to look at the shape of the tomato, as well as an IR view to look for mould or rotting tomato's. We are a machine vision integrator in the USA, if you would like to talk more, please feel free to email me or call. Good luck!

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
+1 (858) 597 1960

Dear Mr. Lee, as Mr. Long said, you would probably need the use of a 3D library like the one we provide: SAL3D. We have already developed quite a similar machine which was used for coconuts sorting based on their size (a video is available at http://www.aqsense.com/applications.html#shell ) You can contact us at dsabate@aqsense.com and we will estimate which are the tools you would need for this development.

Answered by:
David Sabaté from AQSENSE S.L. - Marketing
dsabate [at] aqsense [dot] com
34 972 183 215

I am looking for a vision system to measure inside cylinder diameter at 2 sections. Outside diameter is 2.5 inch and the two sections are located at 0,5 inch and 1 inch from the higher cylinder point. No restrictions about the vision system is made, every solution will be evaluated.

Dear Philippe, after considering your applications with the technical department, we think that we could approach it, using laser triangulation technology together with SAL3D, in two different ways, depending on the requirements. First option would be a lineal scan (as we have already done in some projects where pipes were scanned) This is an easier system but some parts of the cylinder inside would not be scanned. It would be a good option if only the inside diameters had to be calculated or if the part to be inspected was on a know position. Second option would be a rotating camera-laser system next to the cylinder hole. This system would reconstruct the whole inside of the cylinder and therefor any part could be analyzed. If you could write us providing more details on the project (pictures are usually very useful) we could give you a better estimation on how your requirements could be fulfilled. Thanks.

Answered by:
David Sabaté from AQSENSE S.L. - Marketing
dsabate [at] aqsense [dot] com
34 972 183 215

We want to build from scratch a vision system to check error in positioning of robot arms in space. Is there a guideline (books, websites) to follow? Is it better to use a structured laser light source or not? Thanks in advance.

Dear Paolo, to check robots position a stereo system is the typical solution. Usually 3 reference points (like LEDS) are positioned above the robot arm so that using 2 cameras, the exact position can be tracked. If you wanted more precision, however, a laser light system could be used together with a 3D library which could match cloud of points, like the one we develop: SAL3D. If you want more information about this second option, please contact us.

Answered by:
David Sabaté from AQSENSE S.L. - Marketing
dsabate [at] aqsense [dot] com
34 972 183 215

I'm looking for an embedded, self-contained, and low-cost vision system that needs to detect the presence or absence of ~1 mm holes in a surface. The holes will appear black against a lighter background. If the holes are of a wrong diameter outside of some tolerance (about 10%), or are too oblong or are shaped strangely, they should be marked as errors. The max velocity of the moving surface is 1/2 m/s, and a hole could be in a several mm square area. What self-contained vision system could do this? Thank you for your help.

Paul: The PPT VISION M40 Embedded Vision System could solve your problem. See www.pptvision.com for more information. Our software contains very flexible blog detection tools that could be programmed to recognize your defects and flag them. How can we connect by phone or by email to have a more in-depth discussion?

Answered by:
David Wick from Datalogic Automation - Vice President Sales & Marketing
davidw [at] pptvision [dot] com
(952) 996-9500

Paul: In addition to the blob tool that David suggested, PPT has a Circle Gauge tool that may be useful for characterizing the geometric defects of the holes. I can assist you with selecting a machine vision system that will meet your needs. Please contact me offline to discuss your application further.

Answered by:
Scott Gibbs from Mercator Robotec Inc - CEO
scott [at] mercatorrobotec [dot] com
(519) 884-7880

HI Paul ISG has some smart cameras that could perform this function. Give me a call if you want to discuss them. Thanks Kerry

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

Hi Paul - I can provide a fixed cost solution to this application. Contact me directly and I will be happy to help. Thanks, David Wyatt AIA Certified Vision Professional Advanced Level

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

Hi Paul, PPT's smart cameras may be a good choice. Contact me and I'll put you in touch with one of our partners. Here's some "back of the envelope calculations" about the details you have shared: From metrology best practices, we know that it is good to have a 10 times precision tolerance ratio, so if you have a 1 mm hole and you want to hold +/- 0.1 mm on the tolerance band, we'd recommend a pixel size of 0.02 mm. We'll be conservative and disregard that fact that subpixel image processing will get us much better results than this. Our standard resolution camera is a VGA resolution camera (640x480 pixels), if we start with the pixel size of 0.02 mm, the image height comes out to 9.6 mm, which is quite a bit better than a "few millimeters". Now, in regards to line speed, 500 mm / second, and the pixel size we have mentioned of 0.02 mm, the exposure time to get less than 1 pixel of blur should be about 380 microseconds. Backlighting with off the shelf hardware should be fine.

Answered by:
Steve Maves from Datalogic Automation - Manager of Application Engineering
stevem [at] pptvision [dot] com
(952) 996-9504

We are a commercial printing company looking to inspect printed materials off-line. The system would need to compare the printed peice to a approved master and look for defects in the ink areas, voids, spots etc. not color. It would then need to trip a mechanism to deflect the bad product and allow acceptable product to continue. We would create the feeding, shuttle and delivery mechanism however need the camera inspection portion. Our press manufacturers offer these systems during the print process however it is our intention to do it off line.

We can help you with this application. Please contact me directly so that we can exchange more information.

Answered by:
Rex Lee from Pyramid Imaging, LLC - CEO/President
rlee [at] pyramidimaging [dot] com
(813) 984-0125

I can help you with your application. Please contact me directly.

Answered by:
Scott Gibbs from Mercator Robotec Inc - CEO
scott [at] mercatorrobotec [dot] com
(519) 884-7880

Hello Craig, We have a lot of experience in this type of work and can help. Contact me via email and we will prepare a quote. Thanks, David Wyatt AIA Certified Vision Professional Advanced Level

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

Is there a machine vision system/company that specializes in tracking people? There are several conventional RGB systems and far IR for tracking people but is there one using B/W machine vision? I'm looking for a low cost person tracker that can be mounted above an area about 100 foot by 100 foot.

Companies like Tyco Retail Solutions have such hardware and software but they generally do not sell it alone. In the case of Tyco, it is provided as part of retail monitoring systems from ADT. One of our principals, Quest innovations, has developed a software package for this purpose. it was also intended for tracking people around stores but could be used in other environments. If you would like to investigate this, please send me an e-mail.

Answered by:
David Gilblom from Alternative Vision Corporation - President
dave [at] alt-vision [dot] com
(520) 615-4073

I am looking for a vision system (cameras, stands, software and PC, etc.) that can be used both for teaching and research.

Dear Hamid! We can put together a vision system to meet your needs. We have a two tier approach, in that you can use the system for education using a configurable machine vision software and for more involved research applications you can program custom algorithms, if you so desire. We will have to work out the details of what your exact requirements are. Please contact me at your earliest convenience.

Answered by:
Markus Tarin from moviMED - President and CEO
m [dot] tarin [at] movimed.com
(949) 699-6600

Hello Hamid - I suggest you look at Matrox and Dalsa, both fine Canadian companies that would most likely donate frame grabbers and software to the University. Automation Doctor donated our spares to Purdue University in the past, so we are fresh out. Lighting will be the most important thing to get. Try CCS America and Advanced Illumination. Be sure you get menu driven software as well as .NET capabilities. Cognex Vision Pro is a good choice also, and they have free 60 day trials that I am sure they will extend longer for Universities. Vision Pro works with .bmp images so, really, you don't even need a machine vision camera to teach students. Any series of images of the correct size format will do for teaching purposes. Take pictures from any camera (still, camcorder or iphone) and convert them in a PC program to the proper format and load them into the vision software and you are home free. Literally. Good Luck, David

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

Dear Hamid: PPT VISION has provided vision systems to many universities for their teaching and research. We are glad to help you as well. Our local partner Scott Gibbs of PPT Canada (519-884-7880, sgibbs@pptvisioncanada.com can help you define and specify a solution to meet your budget and your teaching needs.

Answered by:
David Wick from Datalogic Automation - Vice President Sales & Marketing
davidw [at] pptvision [dot] com
(952) 996-9500

I am looking to get in into a vision system solution. We are an integrator and also end user. Our factory produces plastic tubes, an aluminum tube, and tin cans. Currently in our company we are using charlock software, pressco , ibea, and sick solutions. I need a simple solution which can be implemented into our production lines very easily. I need to decide which standard is best for our company as I see that there are a lot of good companies selling the vision system. How can I decide what is right for us? BR Tzvi Yeshurun

We provide high speed video recording software that supports cameras ranging from 30 to 2,000 fps at resolution of vga up to 4k x 4k. With our software you can integrate up to 100 different types of cameras.

Answered by:
Luc Nocente from NorPix, Inc. - President
ln [at] norpix [dot] com
(514) 824-2517

Dear Mr. Yeshurun, At AQSENSE we develop a 3D machine vision library (SAL3D) that may be of your interest. We are currently working with BIP (Barcelona institute of packaging) to explore its application for cans and bottles inspection and, among others, It has also been used for hot metal tubes scanning (using a high potency laser with a filter) You can check its uses and specifications at our webpage. As you will see, SAL3D provides the best performance of the market allowing 100% inline 3D inspection. Moreover, we allows exporting the cloud of points to 2.5D zmap representations allowing the integration of classical 2D tools for 3D measures. We also offers help to integrate the library in your product lines thus, making it fast and easy. Please contact me if you think that it may be of your interest and we can arrange a call to further discuss how we could collaborate.

Answered by:
David Sabaté from AQSENSE S.L. - Marketing
dsabate [at] aqsense [dot] com
34 972 183 215

Kibbutz, First please list what it is you need to inspect. 2D or 3D? Stationary or fast? Largest size images and smallest size features to detect. Also, do you want menu driven solutions or can you write in C? I would look at Cognex, Matrox and Dalsa to start with. If these main players do not meet your needs, then there are hundreds of other smaller players that specialize. Your requirements list will guide you. Just a passing note, I used to work with Optrotech and Orbot before they merged when I was the Machine Intelligence Chairman of GM in the 1980s. Both fine Israeli firms if I recall. Good Luck

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

Mr. Ysehurun: The best way to decide what is best for you is to (i) define your goals and then (ii) compare solutions from several vendors that can meet your goals (both technical and business). We have done many projects for customers with plastics and metal forming applications. Our local partner 1Vision in Israel can help you review your goals and then define a good solution for you.

Answered by:
David Wick from Datalogic Automation - Vice President Sales & Marketing
davidw [at] pptvision [dot] com
(952) 996-9500

I am looking for talented vision system engineering resources to supplement our current staff. Can you help identify vision system engineering firms that will work on a contingent engineering (T&M) basis? My company is based in Corning, NY but we have facilities worldwide.

Edward, We are a machine Vision Integrator HQ'd in San Diego, CA, with a secondary office in Oxford UK. What are you looking to do? I can discuss with you our skill sets and see if there is a fit. This probably makes most sense over the phone. Please feel free to contact me via the contact information on this web site.

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
888-508-7355

Mr. Schmidlin, Terus Technologies is machine vision and robotics integrator located in Greensboro, NC. We have well over twenty years of machine vision experience and have integrated vision systems for several Corning facilities in North Carolina over the years. We can work on a T&M basis and can also provide turnkey systems. Please feel free to call or email me for additional information.

Answered by:
Perry Cornelius from Terus Technologies - Advanced Systems Director
perry [at] terustech [dot] com
(336) 339-4508

Dear Edward, My company provides custom imaging solutions on a T&M basis. We are located in Irvine, CA and have an East Coast Office near Boston, MA. We specialize in 2D, 3D as well as thermal imaging applications. In fact we have looked at an application for Owens/Corning in Ohio before. It was a thermal inspection for a fiber glass web application. Please see our web pages for more information. I look forward to discussing details with you in person.

Answered by:
Markus Tarin from moviMED - President and CEO
m [dot] tarin [at] movimed.com
(949) 699-6600

Ed - Check out the list of certified vision professionals on this site. There are 2 levels, Basic and Advanced http://www.machinevisiononline.org/mvo-content.cfm?id=152.

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

Edward: PPT VISION has 30 years of experience in working with Fortune 50 end users like Corning to define and implement machine vision systems. We often provide the service of T&M engineering to end users, and welcome the opportunity to do it for you. Our team of Applications Engineers can help with requirements review, vision programming, component selection, and installations. Please call me and we can discuss your specific needs by phone.

Answered by:
David Wick from Datalogic Automation - Vice President Sales & Marketing
davidw [at] pptvision [dot] com
(952) 996-9500

Edward, I am interested to hear more about what your needs are and how we can be of service to you. Please let me know when it is a good time to chat about what you see this engineer doing and how they are going to help your existing team.

Answered by:
William Corns from Wilco Imaging, Inc. - CEO
wcorns [at] wilcoimaging [dot] com
858-699-2518

Edward, we have done some work at the Corning plant in Guelph, Ontario in the past. I would be interested in discussing opportunities for collaboration with your engineering team.

Answered by:
Scott Gibbs from Mercator Robotec Inc - CEO
scott [at] mercatorrobotec [dot] com
(519) 884-7880

What kind of system would be good for a dynamic environment?

The answer would depend on what is changing in the environment.If you could please provide more details, I may be able to help.

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

Hi, I have a video board which outputs RGB / Ycrcb raw video stream (or any similar formats, such as ITU-R-656) and need it to be connected to a GigE Ethernet in GigE vision interface protocol. Is there any standard off-the-shelf PCB or device that could perform the function, including the software on the PC terminal?

Contact "Pleora Technologies". They might be able to help you on this.

Answered by:
Kasra Ravanbakhsh from Kasra Hooshmand Engineering Co., P.J.S. (KDI) - Managing Director
kasra [at] kdi-co [dot] com
98 21 8867 8023

GigeVision cameras all work with standard off-the-shelf Gigabit Ethernet ports, which can be found at any electronics store like Fry's or Best Buy. You can also buy them from many camera vendors. Regarding the software to acquire from those cameras, that is usually purchased from machine vision software companies. I know there are many. The one I'm most familiar with is the Vision Acquisition Software from National Instruments. It includes something called IMAQdx, which supports both FireWire and GigE Vision cameras. It supports LabVIEW, C++, Visual Studio, and .NET software interfaces. I hope this helps!

Answered by:
Matthew Slaughter from Cyth Systems, Inc. - Vision Specialist
matt [dot] slaughter [at] cyth.com
(888) 508-7355

We are looking into the possibility of a vision system to grade blank rifle stocks on grain size and pattern, along with typical wood defects like knots and voids. Would you have a vision system that could be adapted for this application?

Hello John, if you have enough pixel density and a camera that can provide consistent color characteristics, your project then needs to focus on finding the proper Machine Vision software that is capable of identifying grain size and patterns in the known good or known bad categories. Please contact me for a name of an Lumenera partners in your territory. Thanks.

Answered by:
Corey Fellows from Lumenera Corporation - Director of Business Development
corey [dot] fellows [at] lumenera.com
(613) 736-4077, x184

John, If you have a good idea of what you need you can send us either some stocks for a visual inspection in our machine vision lab, or some images, and we can undertake a small vision study and tell you the results you are likely to get. This would be prior to any contractual decision, as we want to know that we can do what you want, and I'm assuming you want to gain confidence in the company that you might use. We can look at a lot of the information , the real key is describing the groups. My contact information follows, feel free to call me or email me.

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
(888) 508 7355

John - Grain size and pattern are probably not too difficult. The area of concern is a void or knot. A void is a 3D problem. Be very careful if you use a camera (2D Sensor) to solve a 3D problem over a complex surface. It can be done, but I assume there are very interesting cut outs, knobs, nooks and crannies to your stocks upon which you also want to find voids and knots. If you have a specification, we will be happy to evaluate it for you or even help you write it. Please send us your requirements to see if we can help further.

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

John: PPT VISION has experience using our M Series Embedded Vision System in evaluating and grading wooden objects for size and defects. We'd like a follow up phone call with you to better understand the specifics of your application, to see if our M Series product and IMPACT(tm) software can be configured to solve your inspection problem. Please call me or email me with times next week when we can talk. I am in North Carolina.

Answered by:
Michael Hanrahan from Datalogic Automation - Business Development Director
mikeh [at] pptvision [dot] com
919-623-1941

What is the principle of automated vision systems? How would the camera move to the center of the picture (in the case of label inspections)? is there any programming to be done on this?

Sinna - The principle of automated vision systems is to use a sensor(s) to gather light, measure the value(s) of the light before or after any mathematical manipulation and then make decision(s) based upon the value(s). Once you acquire (digitize into computer memory) your image for any inspection, you are pretty much stuck with it. You can move the pixels in memory, or move your display frame to make any part of the pixels appear to be in the center of the screen. Most people move their inspection tools to the part of the captured image they want to inspect. This dictates that you have "locator" tools that first locate the area you want to inspect, for instance a label, before running the label inspection algorithm on the located label. If you are doing robot guidance you can move the actuator to get a better shot, then take another picture. Yes programming or menu changes at some level, by someone, will need to be done to make any changes I hope this helps.

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

Sinna muthiah Meiyappan: We are expert in automated vision systems with 30 years experience in providing the software, hardware and applications engineering services. Usually, there is programming to be done to create a custom vision program to solve each customers specific problem. Fortunately, our IMPACT software uses a drag and drop approach that vastly simplifies creating the program. You can download our software free and see for yourself. We use high resolution cameras for label inspection applications so the camera is fized and the label moves by it.

Answered by:
David Colemam from Datalogic Automation - Business Development Director
davec [at] pptvision [dot] com
(952) 996-9500

Hi, question re getting an independent person to review an application we have with a vision system and the manufacture of nicotine patches . Would like to get a person to review the application and see if applicable for what we want it to do and dealing with the natural variability in the process.

Hello Adrian, We have a lot of experience regarding machine vision system integration for the pharmaceutical industry. Please post more details of your project here or contact us directly. We can assist you to find the best solution for your project.

Answered by:
Kasra Ravanbakhsh from Kasra Hooshmand Engineering Co., P.J.S. (KDI) - Managing Director
kasra [at] kdi-co [dot] com
98 21 8867 8023

Adrian, We are a machine vision integration company, so specialize in delivering these systems. Would you require a review of the system from a statistical verification stand point? Send me an email at the address below if you want to discuss further.

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
888 508 7355

Adrian -- I consult in machine vision and, during my 30 years of consulting, frequently perform assessments like you are requesting. I should mention that I do not sell equipment. So, my assessment will be unbiased. Please contact me either by e-mail or by phone to discuss your requirements and how I can help.

Answered by:
Perry West from Automated Vision Systems - President
perry [at] autovis [dot] com
(408) 267-1746

Hello Adrian, I have many years of experience implementing machine vision systems that inspect medical devices. I can review your application either from the USA or in Ireland. Past pharmaceutical customers include Bayer Corp, Siemens Medical, Ethicon (J&J), Sandoz and Abbot Labs. I have experience in FDA validation. Thanks.

Answered by:
David Wyatt from Automation Doctor Incorporated - CEO
david [dot] wyatt [at] automation-doctor.com
(574) 344-7936

Adrian: PPT VISION is in the business of providing machine vision systems to manufacturing companies like GSK, and we do business with your location in Poland already. Part of the solution we offer is an application review. Can you send me additional information on the requirements you have?

Answered by:
Steve Maves from Datalogic Automation - Applications Engineering Manager
stevem [at] pptvision [dot] com
(952) 996-9500

I'm looking for an auto inspection for label defects. Do you have anyone in Singapore I can contact?

Eric: PPT VISION has significant experience in machine vision for label inspection. Our local partner PPT Singapore can help you. Kelvin Ng is the local point of contact at » PPT VISION (S) Pte. Ltd. » 196 Pandan Loop, #07-20 » Pantech Business Hub » Singapore 128384 » Tel: 65-68720898 » Fax: 65-68724214 » kelvinng@pptsing.com

Answered by:
David Wick from Datalogic Automation - Vice President Sales & Marketing
davidw [at] pptvision [dot] com
(952) 996-9500

We are looking for an online Vision System to identify foreign particles/small insects, etc. inside of PET/HDPE Bottles and TIN cans. Please let us know if you have a solution. Thanks!

Mr. Raju, Do you have specifications and detail system requirements for us to examine to respond? We are a system integrator who specializes in providing turnkey system for problems such as yours. Thanks.

Answered by:
Reddiar Anbalagan from Automated Vision, LLC - President
anbu [at] automatedvision [dot] com
(608) 836-9393

S Raju, We are a machine vision integrator based in USA and UK. We would be willing to run your samples (defects) through our vision lab, to undertake a viability report. If you are interested contact me at the information below.

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
+1 (888) 508 7355

PPT VISION has experience in specifying, supplying and installing machine vision systems that identify foreign particles inside of plastic or glass bottles. We welcome the opportunity to discuss the specifics of your applications to understand how we could configure our IMPACT software and smart camera hardware to provide a solution to you. Our local partner in India is » Semitronik Instruments » 17 CD, Archana Ind. Estate » Rakhial Road » Ahmadabad – 380 023 » Tel: 91-79-22741011 » Fax: 91-79-22741 » semiahd@satyam.net.in » http://www.semitronik.com So I encourage you to reach to them, or to me, with more information.

Answered by:
David Wick from Datalogic Automation - Vice President Sales & Marketing
davidw [at] pptvision [dot] com
(952) 996-9500

We are a pharmaceutical company in Saudi Arabia, we are the second company in turnover in the KSA. We are looking forward for (from surface inspection to blister packs) to be integrated in our IMA blistering machine. Your fast reply is appreciated.

Metter Toledo CI-Vision offers a wide range of solutions that would suit your products, ranging from label and print inspections on your products packaging to integrity inspections on your blister packs themselves. We also have much experience in integrating our vision equipment into other on-line equipment, either as part of the commissioning of a new piece of equipment or retro-fitting into existing machinery, and so would be interesting in looking into how we could fulfill your inspection needs by integrating with the blistering equipment. Please contact me if you would like any further details - or alternatively we have a distributor located in the UAE who covers Saudi Arabia for local sales and support, if you would like to discuss the application in person please forward me your contact details and I will ask them to make contact ASAP.

Answered by:
Gemma Duffy from Mettler Toledo CI-Vision - European Sales Manager
gemma [dot] duffy [at] mt.com
44 1763 257919

Dear Sir, We can surely do it for you. Please take a look at our state-of-the-art color blister packaging inspection system, called Simorgh. You might also be interested in our other pharmaceutical vision systems. Please contact us for more information.

Answered by:
Kasra Ravanbakhsh from Kasra Hooshmand Engineering Co., P.J.S. (KDI) - Managing Director
kasra [at] kdi-co [dot] com
98 21 8867 8023

Ossama, We are a machine vision integrator based out of the UK and USA. We would be willing to run your samples (defects) through our machine vision lab to give you a viability report for your samples. Please contact me at the information below if you are interested.

Answered by:
Andrew Long from Cyth Systems, Inc. - CEO
andy [dot] long [at] cyth.com
+1 (888) 508 7355

Ossama: PPT VISION has significant experience in creating machine vision solutions for inspecting blister packs. We are willing to review your requirements and samples in our vision lab to define a solution that could be supplied to you or your local machine builder. Please reply to me on the next steps you'd like to take.

Answered by:
David Wick from Datalogic Automation - Vice President Sales & Marketing
davidw [at] pptvision [dot] com
(952) 996-9500

I'd like to integrate a vision system with a laser hole drilling system; the vision system will need to identify the form and size holes of the order of 500 micron dia +/-50 micron. Any suggestions?!

James, There are a number of ways to undertake the inspection that you mentioned, and I'm afraid I'm going to turn the process that you need to undertake to determine which one is best for you into some questions. The environment is a critical decision making task in determining the technology to use. Can you backlight the part, i.e. allow light to pass through the hole, and use the silhouette as the measurement? Is the hole 90 degrees to the back surface, or could it be drilled at an angle? Would a 3D inspection of the hole edge give you better information than a silhouette? There are many ways to solve your problem, but it all comes down to the environment in your case. Please feel free to email me (below), with pictures of the setup etc, and I'll try and advise more.

Answered by:
Andy Long from Cyth Systems, Inc.
andy [dot] long [at] cyth.com
(888) 508-7355

The specified tolerance, combined with an assumed P/T ratio of 10%, indicate a requirement for 6-sigma measurement precision of +/- 5 microns. Typical real-world vision-based measurement systems can provide measurement precision on the order of FOV/2500. This indicates a FOV for you application of around 25 mm or less. Many factors degrading precision, including part presentation accuracy, feature position repeatability in the FOV, surface quality, and many more, must be individually quantified to yield a real picture of the capability of a proposed system. Something important to remember - a vision system does not measure a part, but rather interprets multi-point data from an image of a shadow of a part. There is much to consider in establishing the real value of a proposed solution to your application. Let me know if I can help further.

Answered by:
Tim White from Mettler Toledo CI-Vision - Vice President, Technology
tim [dot] white [at] mt.com
(630) 446-7761

James -- What suggestions are you looking for? At first glance, this doesn't appear very hard, but I'm sure as more details emerge the challenges you face will become clearer. My first suggestion, whether you integrate the system yourself or have someone else do it, is to write out a good specification. There are resources on the web that can guide you in creating the specification. If you can't find any suitable guides for this, please contact me and I may be able to help.

Answered by:
Perry West from Automated Vision Systems - President
perry [at] autovis [dot] com
(408) 267-1746

Hi James, I used to do lots of laser drilling of holes in silicon. We used a microscope system that we designed and it allowed the laser to travel backwards to the silicon part as well as allow light to travel up to the camera. A camera was used to identify locations prior to drilling. If you are not using something similar to this then I presume you will want some custom optics that allow collecting light from the hole area as well as guiding the laser. There are quite a few software packages available that can do pattern recognition of the hole as well as metrological functions about the hole. We use MVTec's Halcon or DALSA's Sapera for many of our tasks. We can do some software work for you if you'd like. Just send me some more info and we would be happy to assist. Hope this is of help to you.

Answered by:
Rex Lee from Pyramid Imaging, LLC - CEO/President
Dr [dot] Lee [at] pyramidimaging.com
(813) 984-0125

Does anyone make an automated system that can inspect for particulate, cracks and bottle dimensional defects?

Yes many companies have made such systems. ISG has a number of customers who use our cameras for this application. Send me an email and I can make some referrals. Thanks

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

Peter, There are a number of methods for measuring every product out there. We have sold 3D inspection systems to other bottle manufacturers for inspection of the surface, but it all comes down to the following. - what is the defect I am trying to test for? what does it look like, what technology would show it best. - the cracks how big are they are how do they show up? What technology shows the cracks. This is usually undertaken by putting the products into a vision lab, and then trying it out. - Dimensional measurements can usually be undertaken by any software that has the correct camera lens setup, along with calibration. Look for the term telecentric for lens when undertaking measurements. If you need any further info please feel free to email me.

Answered by:
Andy Long from Cyth Systems, Inc.
andy [dot] long [at] cyth.com
(888) 508-7355

Greetings, I have an FPGA board that has JPEG 2000 images stored in its memory and need to send the images to another device that has a Cemara Link standard-based interface. The Camera Link docuement explains the overall communication protocol: for example for a basic interface there are 3 8-bit pixel-data lines and 3 main control signals FVAL, DVAL, and LVAL. In my case, the images that I have are in the following format: meta data + byes of image data. In the case of basic interface, do I just load the image data 3 bytes at a time, put them in parallel on the pixel-data lines while the FVAL, DVAL, LVAL signals are raised until the end of image bytes? Is this the basic idea? Thank you in advance for your time, look forward to your kind reponse.

That is essentially correct. Typically DVAL is ignored unless needed (data that has too slow of a pixel clock for Camera Link, or to enable BINning for example) since LVAL and FVAL are sufficient in most cases. Most CL framegrabbers have an option to recognize or ignore DVAL. Also, of course, you'll need the output to be coming through the National Semiconductor Channel Link chips or equivalent. You could use EDT’s Camera Link Simulator to help you accomplish your goal here. We can output CL data via DMA through our board. We include a utility and its source code (send_tiffs) in our software package that allows you to send TIFFs from the host PC, through the CLS board and out to another CL device. What you would have to do is modify the existing code that reads in the TIFF to read in the JPEG instead. Our API/SDK is free and available on our website as well if you need further functionality. Please contact me if I can be of any further help.

Answered by:
Cliff Hayes from EDT - Field Application Engineer
cliff [at] edt [dot] com
(503) 690-1234

I am looking for a Vision System to locate a small c'bore hole 0.4mm diameter by 0.5mm deep. Needs to be VERY accurate in locating that hole.

HI Nicky We need some more information to give you an accurate answer. What is your field of view ? How far away will you be from the field of view (working distance) ? We have a variety of cameras that may be suitable. Choosing a lens will be important too. Depending upon what you are attempting to do with the information may impact the lens required. For instance, Telecentric lenses may give a better view. Can you describe your whole project in more detail ? Thanks Kerry

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

Nicky, We can certainly help you with this application; however, "Very Accurate" is subjective. Could you send me more information regarding the application and accuracy requirements?

Answered by:
Nicholas Tebeau from LEONI Vision Solutions - Manager Machine Perception
nicholas [dot] tebeau [at] leoni.com
(586) 979-9900

How accurate, exactly? Over what X/Y range? What is the environment? Vibration, dust, temperature swings? The best you can do is to use off-the-shelf telecentric optics, a pure blue LED light source, a 5 meg camera, and a precision optical grating to calibrate the FOV. If it is a specular surface, lighting becomes critical. A flat coaxial light source like a NERLITE DOAL works well. Some experimentation may be required to get optimum contrast. Get very good free advice from an Edmund Optics sales engineer. These guys know their stuff. Typical 6-sigma repeatability obtainable with a single image acquisition is 1/2500 of the FOV width. So a 5 mm FOV can be expected to yield one-shot repeatability on the order of +/- 1 micron. After obtaining the needed repeatability, your precision will depend on your FOV calibration and the stability of the camera mount relative to the hole substrate.

Answered by:
Tim White from Mettler Toledo CI-Vision - Vice President, Technology
tim [dot] white [at] mt.com
(630) 446-7761

We should be able to help you but will need more information. For instance, the total field of view, time available for the analysis, are objects moving, etc. Please feel free to contact us and send more detailed information about your application.

Answered by:
Rex Lee, Ph.D. from Pyramid Imaging, LLC - CEO/President
rlee [at] pyramidimaging [dot] com
(813) 984-0125

Hi Nicky, regardless of which hardware you choose, the trick is to use the knowledge of what you are looking for. In your case I assume the shape of the holes is known. This will help you either using edgefinders around the hole and simple math to achieve subpixel accuracy (often used on irregular shaped holes) or center of gravity of all the pixels inside the hole (perfect for symetrical holes). To achieve maximum accuracy, you want to have as crisp images of where the edges are and as many pixels covering the hole as possible (but still practical). My collegues have therefore given some suggestions like telecentric lenses, what Field of View (FOV) restrictions your application have, etc. The shape and type of the surface is also most important. If the surface is flat, you can get good images with 2D-cameras and for instance darkfield or possibly on-axis illumination. If the surface is bulging or have rising edges close to the holes, you may have to consider a 3D system instead.

Answered by:
K-J Fredriksson from Jessa Engineering - Consultant
kj [dot] fredriksson [at] home.se
(612) 284-4432

As the problem you describe is obviously key to the vision system being successful, you need to get the sample scanned, and ensure you get the results you expect. Send us a sample, we'll scan it and send you a report.

Answered by:
Andrew Long from Cyth Systems, Inc. - Senior Director
andy [dot] long [at] cyth.com
(888) 508 7355

Hello, I am looking into designing a table hockey based robot game, where vision and control would be used to track the movement of the puck and the response from the robots hockey players. It would need to be able to handle 6 players, including the goalie. Do you know of any cameras and motion controllers that could handle the speed of this type of robot game? Cost is unimportant as I just want to make it work for now. I am not sure if a micrcontroller could handle the speed of the system or if I will have to use a computer. Your help is appreciated. Thanks. Ken Walker Pokanoket Robotic Games

Ken, There are several manufacturers who make high-speed motion cameras -- Princeton Instruments / IDT's M2 and M5, Basler's A504K and Mikrotron's EOSENS cameras can all capture upwards of 500 FPS at varying resolutions. I'm not sure what you require for the robotic game but it seems that would do it. These cameras all produce Camera Link output which requires a high speed framegrabber to capture the data. I work for EDT who makes framegrabbers that can capture these speeds , and our PCIe8 DV C-Link board can work with most models of the above cameras. For more see our website.

Answered by:
Randall Henderson from EDT - Senior Software Engineer
randall [at] edt [dot] com
(503) 690-1234

Ken, We have made a number of vision systems that play games. Do a search for Cyth Bot on youtube for more information, and contact us below if you are still looking for help. Andy

Answered by:
Andrew Long from Cyth Systems, Inc. - Senior Director
andy [dot] long [at] cyth.com
(858) 342-6795

Are you aware of a successful machine vision inspection application where the inspected part is imaged while it is on a vibratory linear track?

I have not come across such an application. I would need to know more information about the application to determine how successful you may be. I would assume a high speed global shutter camera may be able to inspect the parts but that depends upon a lot of factors that are unknown to me at this time. You can send me an email with further details if you would like.

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

Festo used to build a line of vibratory feeders in which the parts were imaged by a line scan camera while they moved. The vision detected wrong parts and misorientation. The illumination was of the shadowgraph type. We supplied the cameras when I was at PerkinElmer ten years ago. I'm sure we could provide something similar for you using Tattile equipment.

Answered by:
David Gilblom from Alternative Vision Corporation - President
dave [at] alt-vision [dot] com
(520) 615-4073

Hi Tom! We are up the street from you in Irvine. We have a lot of experience imaging parts under vibration. In fact we have developed a few systems that actually characterizes the vibration itself of parts which are ultrasonically excited. I'd be more than happy to discuss over the phone or in person at your convenience. Best, Markus

Answered by:
Markus Tarin from moviMED - President and CEO
m [dot] tarin [at] movimed.com
(949) 699-6600

Due to the growing percentage of recycled material, we are confronted with impurities in our finished PET bottles. What would you suggest as being a possible solution?

Hi Serge, thank you for your question. Machine Vision has come a long way in the last decade and bottle inspection is a fairly standard vision application, particularly for gross defects, mis-labeling, mis-capping and so forth. Now for inspection of hard to see impurities in clear PET bottles, this is more challenging and will take some experimenting by someone with vision expertise. If you'd like to send a few samples of good and bad parts, we'd be glad to process them in our vision lab at no charge to you, and provide you a report with our results and recommendations.

Answered by:
Kelly Watkins from Cyth Systems, Inc.
Kelly [dot] Watkins [at] cyth.com
(888) 508-7355

Hi Serge, as Kelly has mentioned, bottle inspection is now a standard application for most vision systems. Many PET bottle manufacturing sites will now have vision systems installed inspecting for contamination in the sidewall and base areas (the most common type of contamination would be burnt particles from the feedscrew blown into the subsequent bottle) while also measuring critical neck dimensions to esnure conformity. The type and size of contaminants that we could potentially find in your bottles would depend on the material of both the bottles and contaminants, and the best way to determine this capability would be for us to evaluate some samples of contaminated product in our vision evaluation lab - this will allow us to determine and recommend the best configuration of system to find your specific bottle defects. Please let me know if this would be of interest and we will be happy to help with this application.

Answered by:
Gemma Duffy from Mettler Toledo CI-Vision - European Sales Manager
gemma [dot] duffy [at] mt.com
+44 1763 257 919

We would suggest a quality control on each PET at the exit of the PET machine. Each PET can be oriented and screened by up tp 6 cameras to analize for any contamination on sides and top of each PET. The NOK PETs would be automatically rejected toa dedicated bin. Do not hesitate to contact us if you require our further assistance. Kind regards, Joao Paes de Vasconcelos

Answered by:
Joao Paes de Vasconcelos from TST Technological Solutions - Managing Partner
joaopvasconcelos [at] tst [dot] pt
351 2654 51 16

I have a requirement to test a camera used in our medical imaging systems. The image originates from a 1K x 1K CCD sensor but the actual output of the camera where I need to test and measure is analog. I would like to reconvert the signal to digital and ideally into Camera Link compatible format with minimal affect on signal quality. I would like to use a frame grabber board but I have not seen any that will accept analog input. Can you recommend a solution?

Hello, There are some analog frame grabbers out there that will probably work with your existing camera. I do think it is a good idea if possible and your budget permits for you to maybe replace the camera altogether with a digital solution, which will save you the cost and hassle of a frame grabber. Please feel free to call or email me.

Answered by:
Michelle Butindaro from PixeLINK - National Sales Manager
michelle [dot] butindaro [at] pixelink.com
631-350-1611

Dear Darrell; This is a bit confusing. Please send me an email to discuss this project. From what I gather - which may not be correct - you have a camera that is 1K x 1K and the output is analog. Since it is analog I might assume that it is an LVDS or other type of signaling ? NTSC is analog but is limited to small resolution. Can you describe this output better ? what frame rate? You want to convert this analog output to a digital output - this can be done a number of ways but we need to know more about your analog signal. You state you would ideally like the digital signal to be in Camera Link Format ? Camera Link is a cabling spec and not a formal protocol. Why do you want to use a Frame Grabber ? Why not simply take the digital signals into a host PC via 1394 or USB ? Why do you want to test this camera ? What kind of medical imaging system is using the camera ? My firm would be interested in helping you redesign this medical system.

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

Darrell, I presume you need this because you have some specialty analog camera that can not be replaced with a Camera Link camera? I haven't come across any analog to Camera Link converters but if you really need it to be Camera Link for some reason (would be good to know why) then one possible solution would be to get the data into a computer via an analog input board, then ship it out via EDT's PCI DV CLS board which can take raw or formatted data and send it *out* as Camera Link. That would essentially mean a whole computer is acting as the Analog to Camera Link converter. We also have Camera Link input boards -- check out our website.

Answered by:
Randall Henderson from EDT - Senior Software Engineer
randall [at] edt [dot] com
(503) 690-1234

Hi! We have built custom camera interfaces many times before helping companies convert from either an obsolete legacy camera to a newer version or to adapt to a different system. Since your application is a medical device you may be concerned with implications involving FDA validation. Let me know, if you can't find anything off-the-shelf. We can help with custom design. Best, Markus

Answered by:
Markus Tarin from moviMED - President and CEO
m [dot] tarin [at] movimed.com
(949) 699-6600

Interested in ADTF Expansion Box for capturing video simultaneously from 4 Megapixel cameras

Dear Myles; Can you provide us with a specification of the ADTF Expansion Box ? Automotive Data and Time Triggered Framework = do you have this spec. also ? We would be happy to discuss a project for this with you. Can you provide me with your contact details ? You can learn more about ISG at = http://www.isgchips.com Thank you, Kerry

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

Dear Myles, We design and build custom imaging hardware. Please feel free to contact me to discuss in detail. Best,

Answered by:
Markus Tarin from moviMED - President and CEO
m [dot] tarin [at] movimed.com
(949) 699-6600

We make software that can capture from 4 cameras simultaneously either compressed or uncompressed. Check out www.norpix.com Streampix 5.

Answered by:
Luc Nocente from NorPix, Inc. - President
ln [at] norpix [dot] com
(514) 824-2517

Hi, I need information about some Inspection Machine that can be checked for particles inside alcohol bottles up to 1000 ml. Thanks for your help

Hi I can refer you to a person who can help you. He is currently in Nicaragua on a vision project. Please contact me. Thank you, Kerry

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

Hi, I have looked at something similar in the past. First step is to define particles. Do you have success criteria describing how small the particles are to still be called particles and require detection? I assume the alcohol bottles are transparent and particles are detectable to the naked eye. Feel free to contact me for further assistance. Best Regards, K-J

Answered by:
K-J Fredriksson from Jessa Engineering - Consultant
kj [dot] fredriksson [at] home.se
(612) 284-4432

Please review the equipment on the following links. Contact me if you have any questions. http://www.seidenader.de/www-seidenader/1-Machines/1e-135-V90AVSB-TD.html http://www.seidenader.de/www-seidenader/1-Machines/1e-136-V90T-TD.html

Answered by:
Ron Lawson from Seidenader Vision Inc. - President
rlawson [at] SVResearch [dot] com
(717) 540-0370, x111

We have an automated router that we would like to add machine vision to, which would allow us to display a picture of the board in the machine to an HMI. The user would then be able to click and add graphics to the picture to select where to make the cuts on the board, and our software would then translate that information into coordinates for the drill. The maximum board size is 508mm X 508mm (20"x20"), and the accuracy of the drill is .01mm. So we need an image that has enough resolution to get that same accuracy. Is there a camera that can do that? And would we have to stitch images together or use a line scan camera to get that kind of resolution? Then, the question is how much is that going to cost?

This problem is similar to the semiconductor global alignment problem where the wafers are 8 or 12 inches in diameter but the alignment resolution has to be in the tens of nanometers. It is not possible to resolve this in a single image so (in one implementation) corner references are used for rough alignment and then fine alignment is done with a lower resolution image accurately referenced to the corners. Other schemes can also be applied depending on the accuracy and stability and motion control on the board platform and drill. We can help but this problem is sufficiently installation-specific that it would require some up front analysis consulting.

Answered by:
David Gilblom from Alternative Vision Corporation - President
dave [at] alt-vision [dot] com
(520) 615-4073

Dear Brian; Have you solved this problem yet ? If not please send me an email or call me. Thank you, Kerry kerry@isgchips.com 585-388-5220 x101

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

Hi, this sounds like a typical 3D-vision application. In order to accurately find edges of the board I would use the height information from a high-end 3D camera and send images and coordinates over Gigabit Ethernet to the HMI and machine software. Best Regards, K-J

Answered by:
K-J Fredriksson from Jessa Engineering - Consultant
kj [dot] fredriksson [at] home.se
(612) 284-4432

Hi, I would like to know if you have a list of providers of vision system for PET recycling. I really appreciate your help.

Natalith - Please describe the inspections you need done. CI-Vision does a lot of PET inspection. Thank you. Regards, Tim White

Answered by:
Tim White from Mettler Toledo CI-Vision - Vice President, Technology
tim [dot] white [at] mt.com
(630) 446-7761

We are in urgent need for a machine vision system to detect a foreign material particles of size ~10 microns and bigger on flat dark surface ~ 250mm x 150mm. We need to be able to detect these particles and remove them before next operation can commence. High volume production line of consumer electronics products Can someone recommend such system/supplier. Some level of customization might be required. Quite urgent request and implementation.

We may be able to help you with your project. Please feel free to contact me so that we can get more details.

Answered by:
Rex Lee from Pyramid Imaging, LLC - CEO/President
rlee [at] pyramidimaging [dot] com
(813) 984-0125

Hello George, I am sure we have a camera solution to fit your application needs. For something like this I would like to direct you to one of our Integrators or depending on the complexity of the job, we could try and work with you directly. Please contact me or send me your contact information and I will call you ASAP. Thanks!

Answered by:
Michelle Butindaro from PixeLINK - National Sales Manager
michelle [dot] butindaro [at] pixelink.com
631-350-1611

This sounds like a typical 2D vision application using dark field illumination. I would recommend Cyth Systems that are in your area. I have worked with them in the past and I have no doubt in their capability to assist you in this matter.

Answered by:
K-J Fredriksson from Jessa Engineering - Consultant
kj [dot] fredriksson [at] home.se
(612) 284-4432

George, We are a machine vision and robotics integrator based in San Diego California. If you would like I could set up a phone call to discuss your problem, and a potential solution. We would need to get a little more into the details of line speeds, field of view, etc. If you would like to discuss further my contact details are attached.

Answered by:
Andrew Long from Cyth Systems, Inc. - Senior Director
andy [dot] long [at] cyth.com
(888) 508 7355

Is there a currently a Vision system that can measure the position of multiple "screw" studs welded to an auto body to verify that they are within a set tolerance?? I am currently tasked with an issue in our Body shop that I think could greatly benefit from a system like this. We currently only measure a handful of studs that we have deemed "critical fit" studs, but when we do have a "non critical" stud it takes us days to catch back up ( ie part fit issues in our Assembly Shop). I would like to be able to measure each stud more often to stop the issue before it can amount to anything. I do not think that checking 420 studs on each and every body makes sense, but perhaps 10 to 20 per body could work.

Dear Alex; I am sure that a vision integrator could help you design a system quickly and at reasonable cost. I refer you to TestWyse LLC. Patrick Wyse could assist you with this project. labviewfl@gmail.com I am sure he could assist you and probably be able to check all the studs automatically. ISG supplies cameras to TestWyse. Regards, Kerry

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

Hello, to correctly answer your question if vision can "measure the position of multiple 'screw' studs" and compare to tolerances, there is an essential piece of information required. The answer is yes if there is either A) contrast between studs and surrounding surface (or such contrast can be created reliably by creative lighting techniques) or B) height difference between studs, screw head tool holes, etc and the surrounding surface. In practicality you also have to consider how large an area needs to be inspected in relation to the size of the tolerances. This ratio usually has some impact on the cost of the inspection system. Feel free to contact me for further discussion on the subject.

Answered by:
K-J Fredriksson from Jessa Engineering - Consultant
kj [dot] fredriksson [at] home.se
(612) 284-4432

We are looking for a vision system that could perform an API standard check of pipe treads from 2-3/8 " to 3-1/2". Would be interested in any leads as to whom may be able to provide this type of technology

Dear Rick; There are many firms who can help you with this application. ISG supplies cameras that could be used. I recommend Jim Campbell at Viewpoint Systems. He could design and build your system for you. Jim Campbell Viewpoint Systems 800 West Metro Parkway Rochester, New York 14623 Phone = (585) 475-9555 Email = jac@viewpointusa.com Good luck.

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

I need a vision system to detect opaque objects from non opaque objects. Basically l need to sort out ceramics from broken glass bottles. So l need a system capable of tracking objects that light cannot penetrate down a 30 degree slide and then ejecting those objects using air valves. So the Camera has to be able to find , scan and eject hundreds of peices of ceramic per hour at very high speeds.

I believe this can be accomplished. Do you have any more detailed specifications ? How many of these systems will you need ? Or do you intend to sell such a solution to your customer base? Thank you, Kerry

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

I am searching for commercially available technologies dealing with optical location/orientation. Must be low cost and provide intelligent localization and mapping. I would like to use this type of optical system for determining changes in orientation of a device. Do you know of any suitable technologies?

We could help you do this. However we would need more information and technical specifications to determine an optimal solution. Cost targets would be nice too. Is this an OEM application ? Thank you, Kerry

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

How can I inspect any object from inside? We have a product which needs to be checked for its defects from inside. How do we check the dimensions of any object?

More information is needed such as how the inside is currently inspected without machine vision, what is the size of the article, what the inspection needs to check, and so forth. Without that information, I might suggest you consider x-ray imaging. The question about how you check the dimensions of any object is far too broad. There are many imaging and image anaylysis techniques that can be applied to measurements. Object size, shape, and material make a difference as well as what characteristics are to be measured. The measurement precision required also affects the techniques. If you have a more specific example, then it might be possible to provide you with information.

Answered by:
Perry West from Automated Vision Systems - President
perry [at] autovis [dot] com
(408) 267-1746

Hello Prasad Dixit, I have some experience of surface inspection and measurements of cylindrical objects where both inside and outside had to be inspected for surface defects as well as dimensional measurements of details. I also have designed several inspection systems looking inside machine parts too small to host the required camera equipment. Let me first assure you that inside inspections and measurements are usually vastly more difficult than outside measurements. This is due to the space constrains: the space inside the object is rarely enough for the equipment itself and the required stand-off distance. As a result you end up with one of the following options: 1. You look into the object from the outside through an opening. This option often puts constraints on the quality of the measurements and inspections. In addition, the size of the opening limits the size of the inside area possible to inspect to sufficient degree. As an example, the larger diameter of a cylinder, ...

Answered by:
K-J Fredriksson from Jessa Engineering - Consultant
kj [dot] fredriksson [at] home.se
(612) 284-4432

Looking to inspect and reject corrugated cases moving on a conveyor at 40 cs/min. Inspection parameters would be flaps tucked, top seals intact, and contents protruding from case.

Hello Steve, This sounds like a pretty straight forward 3D vision application which could alternatively be solved using a multicamera 2D approach. The speed does seem reasonable for finding and measuring the exact position of flaps as well as detecting objects protruding from the case. The only concern from my understanding, is what the seals look like (hopefully not transparent). Any chance you could provide some description or image of what the seals look like, so that I can advice on approach or next steps toward a sufficient solution? Best Regards, K-J Fredriksson

Answered by:
K-J Fredriksson from Jessa Engineering - Consultant
kj [dot] fredriksson [at] home.se
(612) 284-4432

Hello Steve - I think I have a solution for you based on your brief description. We have a complete turnkey vision system unit that does a very similar inspection to what you are describing. It currently looks for missing units, missing caps, different color caps, and debris on caps. One of our client company's has utilized this unit both before they shrink wrap their cases and after to detect for errors during the process. Another utilizes this unit to inspect their aerosol cans in the case for missing caps, cocked caps that would prevent the case from sealing correctly, and missing cans before sealing. This unit is very flexible and adaptable, and I'm sure we could solve your issue utilizing its technology. However based on your short description, I would have to get more information about your cases and process to say for sure. Please contact me for more info and I'll be happy to assist you.

Answered by:
Jeff Griswold from Silgan Vision
jeff [at] mcsvision [dot] com
(630) 790-4497

We want to scan the uneven edges of boards sawn from treetrunk prior to edge straightening and do it in an industrial process. We need to get manageable data on the edge curvature from both sides, the up and down side of the board. Can you suggest any good solutions? Yours, Hannes Tarn

Hello, Wood inspection of boards have successfully been done in industrial processes from upper end sawmills to window frame manufacturers and barrel makers for over 25 years using triangulation based 3D vision. There are several hardware vendors on the market as well as a few good systems integrators with vast experience of this field. I have worked on a number of similar applications myself and would like to know more about your application to point you in the right direction. Best Regards, K-J Fredriksson

Answered by:
K-J Fredriksson from Jessa Engineering - Consultant
kj [dot] fredriksson [at] home.se
(612) 284-4432

I suggest you contact = Amir Pirzadeh President Tisfoon Ulterior Systems 1020 Southhill Dr. Suite 130 Cary, NC 27513 Phone = (919) 677-3060 Email = amir@Tisfoon.com http://www.Tisfoon.com

Answered by:
Kerry Van Iseghem from Imaging Solutions Group - Founder
kerry [at] isgchips [dot] com
(585) 388-5220, x101

We have a new version of Software that can record from line scan cameras. http://www.norpix.com/products/streampix5/streampix5.php

Answered by:
Luc Nocente from NorPix, Inc. - President
ln [at] norpix [dot] com
(514) 824-2517

I'm looking for a solution to prevent from scrap generation in a machining operation. We are talking about detecting variations in the range of 0.100 mm (.004"), or less, from part to part in several locations of the part. It would be also optimum if the detection take place during the machinig process. Do you think that any of the Vision Systems could satisfy our need? Thanks. Adalberto.

Hello, To generalize: detecting variations in the range you describe is very possible at the normal rate in most machine processes (<100 parts / min) However, the shape and material of the part in combination with the dimensions to make measurements in along with space contraints and a few other potential show stoppers could make it more or less difficult. In short, I believe THE RIGHT machine vision system for YOUR SPECIFIC application has very good potential to satisfy your need. Best Regards, K-J Fredriksson

Answered by:
K-J Fredriksson from Jessa Engineering - Consultant
kj [dot] fredriksson [at] home.se
(612) 284-4432

We are looking for some camera solutions to monitor our shop floor CNC systems. Our requirements are: 1) To view machine (spindle) when it's stopped. 2) Be able to view from the internet. 3) Possible works with third party to send e-mail/text messaging. Also, I was wondering if having a single dome-type is better or dedicated to each machine (mounted). If anyone could offer some suggestions I would appreciate it. Thanks Hugh

We have high speed video recording software that would allow you to do the monitoring along with sending the data to the internet for viewing.

Answered by:
Luc Nocente from NorPix, Inc. - President
ln [at] norpix [dot] com
(514) 824-2517

We are using a Keyence vision system with a color camera to detect 2" x 3" plastic parts of many different colors. What would you recommend for a background color? I was thinking flat black.

Hello, In general the best background color would be the one with most contrast to the worst case scenario part, i.e. the part that has the least contrast to the background. For example, if you make various color parts except you never have black parts and very few dark green, dark blue and dark brown parts - then you should should have a black background. If, on the other hand, you often have dark parts, but no white and very few light gray, light yellow, and light pink parts - then a white background is better. Now to the specifics - since you have chosen a color sensor, which is less dependent of contrast once the part is located, the purpose of the application has to be known to give you correct advice. If, for instance, the application require further inspection that is the same for all parts regardless of color, then you should consider 3D vision rather than color vision. If, on the other hand, the application is to sort parts depending on color, the color sensor need...

Answered by:
K-J Fredriksson from Jessa Engineering - Consultant
kj [dot] fredriksson [at] home.se
(612) 284-4432

I have a Motoman robot. I need to install a vision camera to make the robot understand the parts center points. Is there any economical way to do it?

Hi; Your question is about economy and cost, not about technology. I assume then that you already know that machine vision is fully capable of providing both 2D and if needed 3D point coordinates to a robot. You don't describe your parts and process, but those things have everything to do with the success of and the cost of your application. Ultimately though, economy is just a relative question based upon the value of the technology to your process. (if for example the technology costs $100,000, but saves you $120,000 in six months then it could be considered "economical".) I recommend that you thoroughly evaluate the potential return on investment (ROI) related to including machine vision technology in your application. You can analyze savings in things like improved throughput, reduced labor cost, quality improvements and so on. Then you can seek proposals for implementation of the machine vision technology and determine whether it is economical or not.

Answered by:
David Dechow from Aptúra Machine Vision Solutions - President
ddechow [at] aptura [dot] com
(517) 272-7820, x11

Motoman sells a vision system for their robot. I recommend you check with them to see if that vision system will work for what you need. Any other solution will cost far more for the engineering. However, be assured that your requirements can be met with a vision system.

Answered by:
Perry West from Automated Vision Systems - President
perry [at] autovis [dot] com
(408) 267-1746

Hello, realizing the answer to your question may be too late, I still wanted to give it a shot. I have solved quite a few robotics applications that fit your description using 2D and 3D vision systems. In order to determine the correct answer to your question, however, we need to analyze your definition of economical since it is clearly in the eyes of the beholder. What is economical in one application may not be economical in another. I will be happy to give you my best advice if you still have a need and would like to expand on this further... Best Regards, K-J Fredriksson

Answered by:
K-J Fredriksson from Jessa Engineering - Consultant
kj [dot] fredriksson [at] home.se
(612) 284-4432

Hello, PixeLINK has some low cost solutions for adding cameras to your system, please contact us directly at: http://pixelink.com/contact.asp Thank you,

Answered by:
Greg Shore from PixeLINK - Senior Application Support
greg [dot] shore [at] pixelink.com
(613) 247-1211

Hi We are looking for a vision inspection system to inspect the outside surface of a metal cylinder for defects in the metal, e.g. scratches, laminations, chips, etc. Also we would like to inspect the bottom of the cylinder for the same defects. We are looking to accomplish this task at 1k ppm. Is this too much to ask or is there a system out there to handle this?

Hello Roger; At first glance, your application requirements probably are achievable with machine vision technology. The inspections you describe are of the type that are often implemented using machine vision. The challenge in your application, as it is with many machine vision applications, probably is in part presentation. That is: can the part be moved by automation in front of the machine vision cameras and lights in the correct orientation and at the throughput rate you require. To answer that question we'd need to know a lot more about your part and process. Probably it can be done. In response to your question "is there a system out there to handle this?" - likely there is not, at least not at the moment. Your system, with the automation and machine vision technology, would have to be built as a custom system by a company with expertise in machine vision integration. I recommend that you communicate your application parameters with machine vision vendors and integrators to get their input and, ultimately, proposals for the system you need.

Answered by:
David Dechow from Aptúra Machine Vision Solutions - President
ddechow [at] aptura [dot] com
(517) 272-7820, x11

Roger, From a practical hardware standpoint your application is very doable. There are multicamera systems in the field that have been doing similar inspections for the past couple years. Prior to faster computer bus speeds this would have been very difficult. It is now possible with several camera technologies to move images from at least 4 cameras, 3 for 360 degree side views and one overhead or below at 1k ppm. The devil will be in the details. Lighting (off axis for reflective material), part fixturing and an easy to use operator interface.

Answered by:
James O'Reilly from Silgan Vision - President
jim [at] mcsvision [dot] com
(630) 790-4497

Hello, Because your question was entered without a name it was given lower priority - therefore, the late answer. Regardless, I have some experience of inspecting metal cylinders (cylinder sleves for the automotive industry) for surface defects on both outside, inside, top and bottom as well as nuclear power waste canisters (29ton 6m copper cylinders). There are very capable vision systems out there but your question is not easilly answered other than with an "it depends on how small defects you need to detect"-kind of answer. The throughput question depends on what you define as a system. When I have installed one camera systems based on 3D technology I have often competed with multicamera systems based on 2D technology. With that background it can be argued that a multicamera system where each individual camera is inspecting every other, every third, or as many as needed - part while the other parts in between are inspected in parallel, could handle the inspection speed ...

Answered by:
K-J Fredriksson from Jessa Engineering - Consultant
kj [dot] fredriksson [at] home.se
(612) 284-4432

I have a need for a vision system that has the following requirements. Are there any complete vision systems that will work? -IR capture -Min 30 fps capture rate, preferred 60/120 fps or faster -Capture for 1 min when triggered from a TTL signal -Auto save to hard drive, or other recording device and auto reset for next capture -100 feet working distance between camera and control computer Thanks

Hi; Strictly speaking, your application is not machine vision in that it performs only image acquisition and not image processing and analysis. There is however a whole area of products designed for high-speed image capture and recording which perform the exact task you describe. When you say "IR capture", it is not clear whether you mean an image illuminated with an IR light source or true thermal imaging. If it is the latter, you may need to seek out technology specifically targeted for thermal imaging (e.g. FLIR, Fluke, etc.). In general, you can find recording products with a quick web search on high-speed image recording. For the most part these products won't be found on this machine vision technology site. That said, if you can be satisfied with a 30-60fps capture rate, such a system could be built on a PC platform using components readily available from machine vision product manufacturers. In that case, I'd recommend starting by contacting the manufacturers of machine vision software for PCs for input on how to get started. Good luck with your application. David

Answered by:
David Dechow from Aptúra Machine Vision Solutions - President
ddechow [at] aptura [dot] com
(517) 272-7820, x11

Check out StreamPix from NorPix. We are listed in all machine vision directories and we are strictly speaking machine vision. So David, before you talk research. www.norpix.com We have a solution for you. Check out Troublepix. It does loop recording and supports many types of triggering.

Answered by:
Luc Nocente from NorPix, Inc. - President
ln [at] norpix [dot] com
(514) 824-2517

Mn. customer desires to track square footage of conveyorized parts entering a powder coating booth. Parts vary in length, width, height, shape and size. A machine vision application?? Thank you.

This seems like a generally appropriate machine vision application, if it is constrained to reporting the projected profile of the parts in question. In other words, a 2D representation of the part surface. if the customer wants a full analysis of the surface of the parts and a report of surface area, then it likely becomes a 3D imaging application, which could be significantly more complex. I would suggest that you and the customer prepare a detailed application analysis which specifies the parts, the environment, and the desired data to be extracted. A qualified machine vision systems integrator can then more readily discuss the potential for successfully applying the technology to your application. Good Luck David

Answered by:
David Dechow from Aptúra Machine Vision Solutions - President
ddechow [at] aptura [dot] com
(517) 272-7820, x11

I have a vision system that checks on average 65 inspections per unit in 40 seconds using three different cameras. Most of the inspections are part present checks. The environment remains consistent however the background color can change from unit to unit.What is a reasonable percentage of false failures for me to expect from my integrator?

Maybe it's too late, but it would be better to specify how high a false reject rate your process could withstand, and then determine whether that rate could be technologically achievable given part/process variations. Based upon your description of the application, it's clear that what would be considered a very small false reject rate for individual features in some applications - say 0.05%, would nonetheless result in a failure of more than 3 parts in every hundred for your application. That's one part every 25 minutes or so. This may not be acceptable, but it is possibly realistic. Reduction of the false reject rate will depend upon the severity of the part-to-part variation, and possibly upon the capability of the machine vision components and skills of the integrator. Most machine vision challenges begin with imaging. If possible, minimize the part variations through creative lighting and imaging. Perhaps ultimately, the system design and even components may need to be re-considered. Some integrators providing machine vision solutions offer a performance guarantee which states the achievable false reject/accept rates for the proposed system. Perhaps your application is covered. I hope this helps somewhat. Please feel free to contact me if you would like more input. Good Luck David

Answered by:
David Dechow from Aptúra Machine Vision Solutions - President
ddechow [at] aptura [dot] com
(517) 272-7820, x11

We have a very simple project but have not yet found someone who can do it. Basically we are simply trying to capture presence/absence of a day code, 7 characters, as well as legibility (OCR) on a line going about a foot in 5-7 seconds (slow). We've tried DVT and Keyence without success. Our environment is always changing and that is a complication. My question is, can someone recommend a reputable company that stands behind their product and can get us up and running? Thanks in advance.

Greetings; Not to be flippant, but if it were a "very simple" application then probably everyone indeed would be able to do it! Machine vision is not always as easy as it appears to someone who is new to the technology. Presence/absence and simple reading of well-formed characters, printed in a standard computer-readable font is indeed reasonably easy. (though I suspect your characters do not meet these criteria). Determining "legibility" however, can be quite a complex task. You mention that your environment is always changing, and that certainly significantly increases the risk and complexity of any machine vision project. You may need to try other higher-powered vision platforms. (For example, you mention DVT from Cognex; try their Insight product, or with Keyence the CV3000 series) If you are targeting "smart camera" architecture you might also look at PPT Vision, National Instruments, Sick, Siemens, Sony, etc. There are literally dozens of other machine vision systems that might be a good match - look in the Machine Vision Online buyers guide. You should also consider PC-based options from a qualified integrator. I suspect that part/feature variability is the biggest challenge for your application. It's impossible to make a recommendation without knowing your application details. The best starting point would be to clearly define the requirements of your application, the desired outcome of the inspection, the parts, process and variations that will be found in the application, and compile this information into a project specification. Collect representative samples of the product. Use the specification to seek input from machine vision manufacturers, distributors, and integrators. As for a "company that stands behind their product"; all machine vision product manufacturers are reputable businesses that do stand behind their product. It's not their responsibility, though, to guarantee that their components can be successfully used for each and every task. If you would like a fully guaranteed solution for your application you should work with a qualified machine vision integrator who provides full system performance guarantees. My company (Aptura) and others (you can find them on the Machine Vision Online site) do offer this kind of system guarantee. In my opinion, OCR/OCV/Legibility actually is one of the more difficult machine vision tasks, mostly due to part and feature variation. However it should be an achievable inspection. Good luck in your search, and please feel free to contact me with any specific questions. Regards, David

Answered by:
David Dechow from Aptúra Machine Vision Solutions - President
ddechow [at] aptura [dot] com
(517) 272-7820, x11

Dear friends, You all must have seen a petrol tank of two wheeler motorcycle which is somewhat Contour. I need the robot to pick the petrol tank from fixed position and guide it along the object by maintaining 2 to 4 mm gap between petrol tank and object. I need a guidance system which will ensure the gap. If the gap is more than 4 mm then guidance system should adjust the coordinate of robot to minimise the gap also if gap is below 2 mm it should maximise the gap. Technical specifications: Robot Fanuc M16iB, controller Rj3iB Speed of robot 300mm/sec. Thank you for reading, waiting for your precious suggestions. Thanks and regards Sachinkumar

Hello; What you seem to be describing is a vision guided robotics (VGR) task called visual servoing. Your goal appears to be that of tracking a part while the robot is in motion. This is a very "cutting edge" (to the extent that I would perhaps call it "research") VGR application that is beyond the scope of a quick expert answer. However, there may be other options. The most obvious to me from your description would be to use either 2D or 3D machine vision to "map" in space the target "object" that you mention, then move the tank to a location (within 2-4mm) based upon that mapping without real-time tracking. If possible, you may wish to attend the International Conference on Vision Guided Robotics in Novi, MI this fall to learn more about vision guided robotics. Otherwise, please review static 2D and 3D VGR for your application. You may feel free to contact me directly if you have specific questions. Good Luck David

Answered by:
David Dechow from Aptúra Machine Vision Solutions - President
ddechow [at] aptura [dot] com
(517) 272-7820, x11

We need to find an automated inspection system for one of our products; it is a board (4.25” x 7.75”) with 17 7-segments displays and 14 LEDs, we need to verify for all LEDs colors and for no inverted displays and characters. This system will need to be integrated to a VB6.0 test bench system. If it helps to determinate the best option, I could send a picture of the board and information of the displays and LEDs. Thank you in advance, Alfonso E. Cabrera Test Engineer EPIC Technologies, LLC

This application can be performed with any of several smart camera systems currently on the market. The camera would be programmed to determine the color and orientation of each part. Based upon your needs and knowledge, there are cameras which require programming, and others which merely require proper setup using provided software. For an example, I invite you to examine the Sentinel camera shown on our website, www.webspec.com.

Answered by:
Sarah Lowrey from Webview, Inc. - Operations Manager
slowrey [at] webspec [dot] com
(707) 588-7982

I am a grad student at Portland State University. I am trying to find average prices for complete vision systems. What is the range of prices for these products in the market? Also, if you could help me with some information about the top applications of machine vision systems. My group and I are working on a term project about the machine vision systems market (the attractiveness of this market in the future, total market size,etc). Specifically, we are trying to find out whether these computer systems have applications in different segments besides the manufacturing industry such as military, oil & gas, security. If you do not happen to have the information would you be able to direct me to an expert about the machine vision systems industry? This would be of great help since we do not have the funds to purchase any market analysis report. We are relying on external sources such as interviews with industry experts, articles in the internet and university resources. I will greatly appreciate your assistance. Monica Poveda MSFA Graduate 2009 Portland State University

You are undertaking a challenging project! There are organizations who publish machine vision market studies on a yearly basis and invest months and months of time on each study to come up with viable and accurate market data of the type you describe. (And these folks already have inside knowledge of the marketplace!) There really isn't time or space in this forum to even start with an answer to your questions. In short, machine vision prices vary widely depending upon the application, and the technology is indeed used successfully in military, oil & gas, and security markets. I would suggest that you do what the commercial market study people do: develop a series of questions targeted to gather data appropriate to your study, then start interviewing people in the industry: representatives from key component manufacturers, recognized experts, etc. You can come up with lists of people to talk to by studying websites like this one (Machine Vision Online), looking at association memberships (AIA), and by reading key trade publications (Vision Systems Design, Quality Magazine, Photonics, Advanced Imaging, etc.). Having specific questions will help you keep your interview focused and productive. I hope this helps. Good luck with your project. Regards, David

Answered by:
David Dechow from Aptúra Machine Vision Solutions - President
ddechow [at] aptura [dot] com
(517) 272-7820, x11

Is there a machine vision system that could look at 13in x 13in x .05in metal blanks horizontally stacked and leaning against each other and determine the orientation and angle of the flat face so that a robot could align a vacuum gripper to pick only one blank at a time? Any assistance or leads to a solution would be appreciated. Thank you, Lawrence Tom

At first glance the application you describe appears to be one that could easily be achieved with machine vision. The most obvious solution would be one that uses 3D imaging, possibly with a stereo camera pair and/or structured lighting. Of course, any solutions provider will need much more information before making a final recommendation. Please feel free to contact me or any of the integration companies in the AIA business directory to have someone work with you on an application specification. Good luck, David

Answered by:
David Dechow from Aptúra Machine Vision Solutions - President
ddechow [at] aptura [dot] com
(517) 272-7820, x11

We record structural testing of aircraft components, using several cameras which must frequently be as much as 30 meters apart, and desire to upgrade our existing surveillance-grade multi-camera system to data grade. Is there any single-source supplier who can integrate and guarantee a 4-camera system which will reliably record at 30 FPS from each camera with NO dropped frames, and at least 720x480 resolution? Would GigE be a likely interconnect technology for such a solution?

Please consult our web site at www.norpix.com. We supply and develop softtware and turnkey systems for multiple camera recording systems. We have solutions for either 4 or 8 cameras as you wish. Gige is supported with many cameras.

Answered by:
Luc Nocente from NorPix, Inc. - President
ln [at] norpix [dot] com
(514) 824-2517

Question -- For what time period? If you only need to capture for "seconds" then real-time capture to memory may be enough. If you need to capture for minutes or hours, then real-time to hard drive array might be required.

Answered by:
Charlie Dijak from Epix, Inc.
cad [at] epixinc [dot] com
(847) 465-1818

A manufacturer of candles for the home requests us to apply a label to the "front" of 2-1/2" - 5" glass jars. The front being a reference to the candle jar having 2 wicks (separated by about 1") rather than the usual one wick. We can stop & rotate the jars at any reasonable rotational speed but how to [vision] capture the dual wicks for us to stop rotation at a specific position? And at what rotational speed to stop at specific position +/- 5 degrees? Thanks for reading, Dennis

Hello Dennis; To start, I'll presume that imaging conditions allow a vision system to properly extract the rotation of the two wicks relative to the center of the jar opening. The easiest and most accurate way to guide the rotation would be to capture just one image of the wicks when the jar is in station. From that one image, calculate the rotation needed to properly position the jar, then supply the automation controlling the rotation with that information. (For example, If it is a servo or stepper you may be able to directly send it a movement command or number of steps.) "Tracking" the rotation like you're describing is possible, but will not be as accurate or maintainable due to the latency of image acquisition, processing and signaling. If you are totally dedicated to using this "tracking" method, you will first need to calculate how long your image acquisition and processing will take based upon the algorithms needed to determine the wick positions. Then you will be able to calculate the maximum speed of rotation based upon the estimated delay in getting the "stop rotation" signal. For example, say you have a jar with a circumference of about 7". Through testing you determine that a single vision cycle to find the wick position takes say, 200ms. This means that the part will continue to move for 200ms before the vision system signals it to stop. Calculating for the speed of travel during that time, in this example the circumference of the part can not be moving faster than about 2"/second to hit a +/- 5 degree window. Note that this calculation does not take into account any positional error introduced by the imaging, nor any error in stopping the rotation. Note also that the latency will change every time you change any part of the inspection program. Your numbers will, of course be different, but hopefully this steers you to the right calculations. Good Luck. David

Answered by:
David Dechow from Aptúra Machine Vision Solutions - President
ddechow [at] aptura [dot] com
(517) 272-7820, x11

We are interested in a vision system that will do the following: 1. Detect a paint band on a flat piece of glass and determine if there are any paint voids in it. 2. We would want 0.5 mm recognition. 3. The area would be 27 inches across. 4. Detect characters and measure those character to be in the proper place compared to a found edge. Our images are silkscreened on, so there is some variance in the characters, we understand that. We are concerned with missing letters, letters filled in and so on. We are under the impression a line scan system is what we need and up to 12k cameras. Any suggestions or comments are appreciated.

It is difficult to make specific comments on your application without an overall evaluation. However, in general, the type of inspection you are describing certainly may be done with line scan technology, but might also be accomplished with area scan cameras. The accuracy of the imaging is one consideration, but, by my calculations you will need only about 6K-7K pixels across the 27" part to achieve the 0.5mm feature detection you require. (conservatively - 685mm field of view / approx. 0.1mm desired pixel resolution = 6850 pixels). So, the final solution could be multiple area cameras, or an 8K line scan camera. You will have to weigh other technical factors such as speed of travel, lighting, and image distortion to decide which solution is best. Considerations such as system complexity and usability also could be a factor. Hope this helps. This sounds like a very viable application. Good luck. David

Answered by:
David Dechow from Aptúra Machine Vision Solutions - President
ddechow [at] aptura [dot] com
(517) 272-7820, x11

Ask a Question

View category list

Current Category:
Vision Systems

Click here to view archived questions for this category.

Top of Page

Email This Page

Print This Page

Ask Your Question

Close Without Submitting

Fill out the form below. Required fields are marked with (*).

Choose a Category: *

Type Question: *

First Name: *

Last Name: *

Uncheck box if you want your LAST NAME hidden from our Experts, it will not be displayed on the website either way.

Business Title: (will not be revealed)

Company: *

Uncheck box if you want your COMPANY NAME hidden from our Experts, it will not be displayed on the website either way.

Email: (will not be revealed)

To help our Experts provide the best possible answer, please provide the following information (this will only be revealed to our Experts, it will not display on the website).

My Company Is: *(please select only one)

My Company's Primary Product or Service is: *

My Company is Located:

City: *
State/Province: *	(Type NA if not applicable)
Country: *

I agree that by accepting free resources from AIA that I may be contacted by members of the trade association. AIA sends out periodic correspondences about Association updates and news. This is considered a part of the business transaction you have chosen.