- This article is filed under:
- » View All
Machine Vision in the Wood Products Industry - Part 1 - 3D Based Systems
by Nello Zuech, Contributing Editor - AIA Posted 08/24/2001
Machine vision has been in widespread use in the wood products industry for many years. Systems include those based on 3D to make volume measurements on the wood during various stages of lumber manufacturing as well as those based on color used for grading applications. The first part of this article will concentrate specifically on the 3D-based machine vision systems. Applications of these optimizing techniques include: bucking, primary breakdown, cant, edger and trimmer operations.
The vendors of scanner/optimizer equipment to the sawmill industry can be classified into three generic groups, although some companies cut across the groups. These groups are suppliers of: machine vision/scanners/sensors that are used in scanner/optimizers: sawmill machinery suppliers who build their own scanners/sensors, and system integrators. The integrators are of two types: those that are more into controls and integrate scanners and machine vision and deliver turnkey optimizer systems, and those that are integrators of sawmill machinery including controls and scanner/optimizers.
A questionnaire was emailed to suppliers of volume-based scanner for optimizers used in the wood products industry posing several questions. What follows is the distillation of the responses.
What does the wood products industry want in an optimizer scanner?
Walt Pastorius of Laser Measurement International (LMI), Delta, B.C. Canada, suggests: 'Reliability in the harsh environment is the first request. As the need for higher production increases the industry is requesting faster scan data. For primary (log scanning) the need to scan away from the machinery that is handling the logs is important followed by accurate 3D profiles at approximately 200 to 500 Hz scan rates. For lumber, 1KHz has been the industry norm. This could be doubled to 2Khz to meet most current and future needs. Scan density is less important as the industry is served by 1', 2', 3' and 4' scan densities (along the board) now with the most common being 3' and 4'. Advancements would be more inline with feature extraction from smart systems.'
Yvon Hubert of Comact Optimisation, Inc., Broisband, Quebec, Canada succinctly suggests: 'Optimum recovery, value, flow control for a higher piece count.' While Dan Smith of Porter Engineering, Ltd., Richmond, B.C. Canada adds 'Reliability, correct data generated from scanner, repeatability of data.' And Karl Gunnarsson of Integrated Vision Products (IVP), Woodenville, WA, notes: 'The industry wants to increase the yield and production levels, decrease the number of rejects and have a flexible production with short change over times. The optimizer scanner should, therefore, measure the material with high accuracy (< 1mm) at a high scan rate. This will require an effective frame rate of around 2000 frames/sec in some applications. …. The system should be robust and easy to maintain.'
How important is 3D-based scanning to the wood products industry?
All the respondents indicated that 3D-based scanning is very important and is becoming necessary. As Walt Pastorius observed, 'This has been the most significant advancement in the last 5 years for primary log scanning. It has allowed the more robust software packages to improve the recovery for at least a 3 to 4% and sometimes as much as a 12 - 15% yield improvement.'
What are the advantages?
Dan Smith cites 'Better data defining real shape of wood' as the main advantage. Walt Pastorius adds 'The solution represents the real recovery and can allow for variations in the wood shape. Prior to 3D log scanning the assumption was that the log was round or oval with the resulting solution either wasting wood or the piece having to be remanufactured due to missing wood the scanner could not detect. Additionally log rotation has gained recovery that could not be exploited with 2D scanners.' Karl Gunnarsson also notes 'One big advantage of 3D instead of 2D in log scanners is the treatment of 'out of round' logs. A 2D scanner can't separate a round log from an 'oblong' one.'
Are there some tradeoffs that might favor 2D scanners over 3D scanners?
While all suggested that there are times when only a lower priced system can be purchased (implying that 2D is better than nothing at all), Walt noted that 'Shadow scanners (2D) can also be more robust in tough environments' and Dan observed that '2D scanners are still more appropriate for bucking applications that have logs with bark on them.'
Yvon also suggested 'Sometimes the amount of information is too much to be used by the optimizer, in the bucking, for example. But this is becoming less of a factor because today there is sufficient computing power that can make good use of the 3D data now. Also, in some applications, like presorting, it is not worth it to put a 3D scanner. And last, the mechanical accuracy has to be there to realize the benefits.'
Is this an application-specific issue: bucking, primary breakdown vs. secondary breakdown applications or within the secondary mills even specific to cant, edger, trimmer, planer applications?
Walt of LMI observed '(In) Primary (mills): 3D scanning advantages for solution can be applied to all levels of scanning. The higher cost of 3D scanning for carriage applications has to weigh against the recovery advantages so this is dependent upon the mill size and amount of production. (In) Secondary (mills): In North America the cutting rules allow a percentage of wane (missing wood) allowance on the edges. 2D scanning can miss this or can only assume the edge shape resulting in missed cuts.'
What are the drivers for adoption of 3D scanners in optimizer systems?
William J. Briskey of Lucidyne Technologies, Corvallis, OR, observes, 'Logs are not round. Nor are they straight. When a board is cut from a log the tensions held in different sections of the wood fiber affect the resulting straightness of the board. Additional movement occurs during the drying process. The grade of a board, and thus its value, is determined by how straight it is, missing wood (it is called wane, caused by the board being cut from the outside of the log), visual defects (knots, splits, stains, rot, fiber type, etc.), and strength defects.
All decisions made in the process are aimed at getting as much value from the fiber as possible. However, in general, the expense and state of technology has limited (until recently) the scanning and optimization task to only optimizing the geometric shape of the log, cant, flitch, or board.
In running a sawmill operation you try to cut what the market calls for today with an eye toward tomorrow. Prices change daily, which means in many cases that a mill must constantly retrain their people who make the cutting decisions, Many of these people make less than $10 per hour and there can be a high turnover rate. In many cases they are also pushed to the point where their mistakes are just barely tolerable. This is why automated scanning and optimization is of such interest.'
Dan also noted, 'Four or five years ago it was driven by sales people looking for the next thing to sell. However the initial systems that have been installed have proven that there is an increase in recovery, value and grade.' And Walt suggests, 'Return on investment and getting the best recovery out of the wood. The resource costs are increasing so 3D scanning (and the associated software and handling equipment) can result in a higher return for the same raw material costs.'
What is different today than say five years ago when the technology also held much promise?
Dan notes that the wealth of experience with the technology has resulted increased confidence of the buyers. 'Mills like to see proven things. Very few mills want to be the 'first'. The cost of 3D scanning is much more than 2D scanning and no one could justify payback. Now that systems are installed there are numbers to work with.' William suggests economic drivers 'There is more competition for raw material (trees), the cost of labor is higher, and technology wasn't far enough along to do more than a geometric optimization.'
Karl suggests 'Maturity and experience in the systems. Five years ago there was an over confident approach of what can be accomplished. Many vendors adopted research from universities that was not very well tested. Today the vendors have more experience from the wood industry and often have a solid background in making control system. It's the machine builders that have learned how to build scanners.' And Walt summarizes it by saying 'The 3D scanning for logs has stabilized into a mainstay. The improvements have been in speed and density of data. 5 years ago the scanners were operating at 30 to 60 Hz, now they are operating at 120 to 400 Hz. Remember that the improvements are also due to the software and handling equipment after the 3D scanners data is produced. For board scanning, linear scanning has made slight inroads to the market with the most attention on the high speed lineal scanning after the planner.'
What are some factors that can be used to justify the purchase of a 3D system?
Bottom-line results can be demonstrated by increasing the wood recovered from a log, as well as the value and the grade. As Walt suggests the return-on-investment can be substantial - 'For example a mill the cuts $35M worth of wood per year with an increase in 1% can realize a return of $350,000 per year. Imagine if they achieved a 5% ($1.75M) or greater return.' He also observes a tangential justification factor - 'For the environmentalists, there are less trees being cut for the same production.'
What companies are adopting the technology? Big companies? Medium sized? Small? Is there any other way you might characterize those companies adopting this technology?
Walt observes, 'The early adopters were the big companies due to the higher budgets available. Today in North America, it is a standard product. The small independent may have budget restraints but they know they need the 3D scanning and optimization if they have any chance of competing, especially in the high production mills.' In general, as noted by William, 'Larger companies can better afford the cost and the payback is shorter with the higher volume a larger company puts through their operation.' And Karl suggests that medium sized companies 'that have the competence and resources' are also applying the technology.
Of the applications generally cited: primary breakdown, cant, edger, trimmer, planer, which is the main application for which 3D scanners is being adopted? And Why?
While Dan suggests 'Primary breakdown and curve sawing exclusively ship with 3D scanners now,' Walt offers a somewhat different perspective 'Historically, the edger and trimmer transverse scanning has been used with transverse dot based triangulation scanners. These systems have squeezed as much production as they can out of the wood. There is very little room left based on dimensional scanning. The last 5 years has seen line based 3D camera technology applied to lineal log scanning for primary breakdown. There was much room for improvements in recovery so this is where the most advancement has been made. The more difficult applications were carriage and bucking scanners due to the standoff required by the scanner. The final area being tackled is the lineal board scanning out of the planner. In the last year advancement have been made in the speed of camera technology that is allowing boards to be scanned moving 2000 fpm. The next area to be tackled is lineal grading and this is just starting.'
And Karl, still another 'All can benefit from 3D but the main applications have traditionally been primary breakdown, cant and edgers. The main reason has been that the speed of 3D measurement has been too low (50 profiles/sec). That was OK in log scanning and cant. The edger scanners are normally working in transverse mode and can, therefore, have lower speed. New high-speed 3D scanners make it however possible to build lineal feed edgers with high accuracy. The same possibilities apply for trimmers and planers.'
Where is the major market geographically? United States, Canada, South America, Europe, Pacific Rim, etc? Are there reasons why regions outside of North America have not been as quick to adopt 3D scanner-based optimizers? Are those reasons region specific?
Walt summarizes 'North America has been the most aggressive in adopting 3D technology. This is due to the economies of scale and the need to meet the cutting rules. Australia / New Zealand have had similar requirements but are limited by cost and available resources to support the higher technologies but this is changing. Europe has been more aggressive with visual grade scanning and due to the different cutting rules has not readily applied 3D scanning to board optimization. The log scanning for Austria and Germany have government rules applied to them that have only recently (last 3 to 4 years) been met by some higher speed technology.'
While Karl observes 'United States, Canada and Europe. Others regions may have cheaper labor. Some markets may not have the infrastructure to maintain advanced machinery. The globalization of wood producing has quickly leveled the playing field and we can no longer assume that level of sophistication is different. The pulp market is a good example where the most profitable producers are located in Brazil and Indonesia. We should see a similar development in wood industry. We must of course take into consideration the cost of transporting the products from mill to the consumer. This adds a whole new dimension to the issue.' Karl also observes 'In Scandinavia no mid/big sized sawmills will process wood without a 3D scanner. There in no question in my mind that Scandinavian mills are well ahead of the North American mills when it comes to automation. Within 5 years we predict that most mills will have 3D scanning. Part of the reason for 3D inspection in Scandinavian was traditionally high relative cost of labor (not the case any longer).'
Which segment of the wood product industry is adopting the technology and why? Hardwood? Softwood? Are there other ways to segment the market for 3D-based scanner/optimizers?
Consensus expressed is that since softwood is more abundant and softwood plants are generally larger than hardwood plants, the migration of 3D scanners into the softwood market has been faster than in the hardwood market. Walt suggests, 'Softwood has always had a larger budget for optimization due to the quicker paybacks from higher production.' Another barrier in the hardwood market suggested by Karl is 'For hard wood the cost of raw material and in many cases the requirements for higher accuracy has made it harder to implement a complete inspection system.' Another barrier for the hardwood market suggested by William was 'hardwood mills are generally smaller and have less capital. Technology is thus slower to move into the hardwood producers.'
What makes for a successful installation? What does the buying company have to do to assure the installation will be successful? What does the vendor have to do? What should be avoided to avoid failure?
While 3D scanner technology in wood optimizers is a relatively mature technology and, as noted by William, many of the mills have electrical and mechanical people capable of dealing with these systems, the following summarizes the tips:
- Become educated and have a good understanding of the system, what he can expect from the system
- Be prepared to allot sufficient resources
Plan for success
- Understand customer expectations
- Understand environmental issues
- Sufficiently pretest equipment before installation
Offer sufficient training and insist that training take place
What does the future have in store: What are the technical trends in the wood products industry that will impact the 3D scanner requirements and performance requirements? What are technical trends in the 3D scanner market that will impact future 3D scanner products?
The consensus seems to be that 3D technology is fairly mature and an integral part of the wood business. The future will bring faster systems applying faster computers with higher resolution sensing capabilities yielding higher accuracies. Improved man-machine interfaces are also expected.
Significantly, the future will see the adoption of techniques that can simultaneously perform internal scanning of the log or the lumber. A number of companies are in the early stages of adapting ultrasound and X-Ray techniques to scan for internal defects. This capability provides the complementary capability to grade the log or lumber so optimization of yield becomes based on both volume and grade. This then leads to better value optimization where the most value is derived from the way the log or lumber is cut by incorporating market pricing information into the optimization programs.
My thanks to Walt Pastorius, Karl Gunnarrson, Dan Smith, William Briskey and Yvon Hubert for their input for this article.
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.