- This article is filed under:
- » View All
Machine Vision in Packaging
by Nello Zuech, Contributing Editor - AIA Posted 07/08/2003
While not a market for machine vision as large as the semiconductor or electronic industries, the packaging industries collectively represent the third largest market for machine vision. By collectively I refer to all applications in container inspection as well as packaging line applications in the food and pharmaceutical industries. Given that most packaging is used with consumer goods and virtually all consumer goods packaged for North American consumption is actually packaged in North America, this is a market that is not as subject to migration offshore.
While the machine vision companies that address packaging line applications have not seen the growth experienced by those companies addressing applications in the semiconductor and electronic industries, neither have they seen the roller coaster cycles. For the most part, packaging line applications have seen a steady growth.
While cost may have been a barrier to the widespread adoption of machine systems on packaging lines, the advances in price/performance of machine vision systems has resulted in diminishing the cost argument for not doing anything. Today one finds vision sensors, smart cameras and embedded vision processors that sell for well under $10K. These lend themselves to installing several along the many value adding sites of a packaging line, each performing a single inspection to verify quality following each specific value adding function.
Before starting my consultancy I was with a company that installed one of the first machine vision-based empty bottle inspection systems on a filling line. It seems that at one of the plant sites of the company a mouse had crawled into a bottle while the case of empty bottles was in the warehouse. Unable to get out of the bottle the mouse had died. Although the bottle went through a washer before being filled, the bloated mouse did not come out of the bottle during the wash cycle. A customer who had the misfortune to use the product in the container on her breakfast fare sued the company asking for pain and suffering damages since she would no longer be able to enjoy that breakfast fare ever again without reliving the incident. The President of the company responded quickly and mandated that all such filling lines be equipped with a detection scheme to assure no such incident would ever occur.
While even 20+ years ago machine vision technology existed to address the application, the challenge as much as anything was that never again did a mouse crawl into a bottle. While the system never detected a mouse again, it did experience false rejects. Though not many, when filling 150 bottles per minute they did add up, much to chagrin of the line people. The result was that the line people found the system a nuisance and turned it off. This lead to one of my golden rules when counseling on a prospective machine vision project – will the system address a recurring concern so that its performance can be validated?
One of the earliest applications for machine vision (way back in analog days) was the inspection of returnable glass bottles before being refilled to make sure there were no visual contaminants in the bottom of the bottle and the lip/seal of the bottle was not damaged. Similarly at the glassware manufacturer the technology was adapted to make sure there was no birdswing or glass filaments attached to the bottom or the sidewall. This was then followed by systems to detect both container reliability (cracks, checks, etc.) and cosmetic defects (bubbles, etc.).
Today one finds digital versions of these systems using arrangements of cameras to inspect the entire bottle: sealing surface, neck, sidewall, heel and bottom. While most of these are installed after the glassware has cooled, there are a couple of systems that actually look at the hot gobs to make sure shape is correct before forming the bottle itself as well as read codes on the bottom or heel of a hot bottle to quickly remove problem cavities forming poorly shaped bottles.
At the fillers one will also find a family of machine vision systems. If returnable bottles are being filled (still a practice in Europe and the Pacific Rim) systems inspect them to make sure they are empty, that the seal has not been damaged and the sidewall to make sure that it has not been scuffed too much resulting in shabby appearance. Another application one finds where refillable bottles are used is a sorting application where bottles are sorted by brand. It would not do to have a diet drink in a non-diet drink bottle.
In the case of glass bottles because they are relatively fragile and the manufacturing process yields concerns such as birdswing or glass pieces in the bottle, inspection is critical. In the case of the manufacture of plastic bottles, health concerns to consumers as a result of a failed manufacturing process are not as critical. Hence, while machine vision has been offered to perform essentially the same types of inspection following the manufacture of a glass bottle (sealing area, neck, sidewall, heel and bottom), these systems have not been as widely adopted by the plastic industry. More widely adopted are systems that inspect the plastic preform whose shape and geometric properties are critical to the manufacture of an acceptable bottle. However, as misshapen bottles can create problems at the filler leading to spillage, for example, and corresponding annoying line shut downs to clean up the mess, fillers are putting more pressure on their suppliers or on their own systems for producing bottles. Hence, there is growing interest in inspecting plastic bottles for various concerns.
In the case of plastic and glass bottles one finds machine vision systems used to inspect their closures. These can be for both inside and outside inspection. Inside to guaranty the properties of he seal area as well as for aesthetic concerns and shape and outside for cosmetic concerns as well as deco.
After filling, one finds systems that inspect the labeled product (though in the case of plastic bottles made with an in-mold labeling process the properties of the label are inspected after molding). This inspection not only assures the presence of the label but also that it is the correct label, properly placed, not skewed and without folded corners and other wise cosmetically correct. These systems might even inspect for any overfill on the sidewalls.
Often these same systems post filling can also inspect the closure to make sure it is present and not cocked. Today x-ray-based machine vision systems are available to check the fill height. These systems are replacing gamma ray fill height detectors because unlike the gamma ray source the x-ray source can be turned off when not in use – a workplace safety concern. If the container is transparent or translucent it may be possible that optically-based machine vision systems will be able to perform fill height detection. Some of the x-ray systems are sensitive enough to also detect foreign artifacts in the container, as, for example, a glass particle in a baby food jar.
After filling the bottles are placed in cartons. Machine vision systems can be used to make sure the correct complement of bottles is in the carton and that they are properly positioned. When a sealed carton is to be inspected an x-ray-based machine vision system may be the answer. Other systems are used to inspect the labels on the cartons and to make sure the cartons are properly sealed.
In the case of metal containers one finds systems at the manufacturer that inspects beverage and food cans as well as lids. In the case of the lids inspection involves validating the application of the sealant as well as inspecting for shape and cosmetic concerns. In the case of the container itself, systems validate the shape, inspect the interior to make sure the coating is complete, and inspect the sidewall and bottom for cosmetic concerns. There are also systems that inspect the outside of the can – the deco side. In this case they verify that that the deco is correct, color is correct and pattern does not exhibit cosmetic concerns.
Over the years a number of companies have offered machine vision-based seal integrity systems for the food industry, following can fill. It is not apparent these have been widely adopted. This reminds me of another early project I was involved in before my consultancy. The Army was looking for a system that could inspect the seals on their ready-to-eat retort packages. I turned the project over to our application engineer who was a retired PHD out of one of the major think tanks in the country. The next morning he called me into his lab to give me a demo that he had worked out in his home lab the night before. Turns out the approach he demonstrated was based on acoustic signal processing. I had to remind him that we sold machine vision solutions. Since then several companies have also developed machine vision-based seal inspection systems for plastic pouches and tubs.
Insofar as the labels themselves, machine vision systems can be found at the printers of the labels. One can also find them at the printers of soft packaging materials themselves – paper or plastic wrappers. The type of video-based systems one finds at the printers of consumer packaging range all the way from those that perform mark registration based on reading the marks in the border to those that perform a comprehensive inspection on the entire label or wrapper. Mark registration is critical to assure color registration as well as print-to-print registration.
In the case of inspection one finds simple systems that simply sense when a change has taken place in the print pattern (change could be due to color, pattern shift, etc.) and flags an operator to confirm the system’s observations and take corrective action. One can also find systems based on either color cameras or color line scan cameras to perform untended complete inspection on the printed pattern. These systems can flag conditions that might lead to failure before actual failure is observed and alert the operator to make the necessary changes to assure quality is consistent. Alternatively, they might send a signal to a marker and record the location in the printed product that indicates where a concern condition has been detected.
Throughout the consumer packaging market one finds that printers apply date and lot codes. In some cases these can be applied on the labels immediately before application on the container. In the case of beverage and food cans one often finds them on the bottom. In some food applications of glass and plastic one might find data and lot code applied on the top of or along the or sidewall of the closure or on the label or on the neck or sidewall of the bottle. Today machine vision systems can handle just about every printing modality and perform reliable date and lot code verification without a high incidence of false rejects. Some systems not only perform verification function but also examine the print to assure legibility and that print aesthetics conform to standards.
The pharmaceutical industry, in addition to some of the abovementioned applications, has a couple of other packaging line applications. Blister packs are commonly used to package solid dosages of one type or another (tablet, capsule, gel, etc.). In this case, color-based machine vision systems are generally used to make sure that each blister has a solid dosage, that it is the correct color, correct shape and complete. The sealed package is generally stamped with a data and lot code and that, too, can be inspected by a machine vision system.
Another application found in the pharmaceutical industry is machine vision-based proofreading. Virtually all printed materials for the pharmaceutical industry are critical. They must be guaranteed to match the original design. The inadvertent misplacement of a period (0.01 vs. 0.1) can be disastrous. Materials that must be proofread include labels, cartons, inserts, outserts, etc. In the case of inserts and outserts one find nearly microscopic print. These systems require high-resolution scanners and typically employ precision registration techniques to align the piece under inspection precisely to the stored artwork. Image subtraction will yield those areas in the artwork that differ from the original. An operator generally then verifies whether the highlighted areas identified by the system are true errors or not. These systems are generally used to inspect the first piece of a production run and samples throughout the run.
Suppliers of the machine vision systems for packaging line applications are a mixed bag. In the case where an application requires special material handling to handle the package one will generally find companies that provide application-specific machine vision systems – turnkey fixed solutions that easily integrate into a manufacturing line. Systems that inspect glass, plastic and metal containers and closures or perform online print inspection are examples of application-specific machine vision systems.
Where an application involves the integration into a specific piece of packaging machinery one generally finds that a system integrator using commercial-of-the-shelf machine vision products can provide the solution required. Sometimes the machine vision vendor will provide the machine integration. Today as machine vision technology has become more commonplace, it may even be that the buying company has the integration skills required. Examples of these applications are generally found following filling operations: closure/label inspection, date/lot code, etc. In these cases the machine vision vendor provides ‘‘canned’‘ software and man-machine interface that addresses the specific application.
The value of machine vision systems in packaging operations has been demonstrated repeatedly. Assuring products are labeled and coded correctly has prevented costly recalls and at the same time made recalls more reliable. Assuring the integrity of seals also eliminates the possibility for product contamination and the consequences of such contamination. Ultimately, the use of machine vision on packaging lines assures customers will not be put off by the poor appearance of a package and that they will be satisfied with the contents of the package.
There are currently no comments for this article.
Leave a Comment:
All fields are required, but only your name and comment will be visible (email addresses are kept confidential). Comments are moderated and will not appear immediately. Please no link dropping, no keywords or domains as names; do not spam, and please do not advertise.