Meristem: Wheat Science

'Neural network' technology modeled on the human brain may provide the key to addressing mounting challenges for Canada's grain handling system.

Pressure on Canada's grain handling system has increased dramatically in recent years, as it faces more crops, more varieties, more customers and more specialized food safety and quality concerns than ever before.

Now, a new generation of technology is emerging to meet these new challenges. Among the most promising is Acurum, a new automated grain analysis technology that offers sophisticated, objective and rapid analysis of grain. The technology was developed over seven years at the Agriculture and Agri-Food Canada (AAFC) Lethbridge Research Centre and has been licensed by DuPont Canada Inc., for marketing worldwide.

"One of the challenges facing Canada's grain industry is that the long-time quality assurance system for cereal grains is based on human visual inspection, which is subjective, labour-intensive and carries an inherent risk of inaccuracy and inconsistency," says Dr. Glenn Coulter, Marketing and Licensing Manager for AAFC. "The technology behind Acurum is designed to eliminate these problems - it provides uniform, objective analysis that is much faster and far more sophisticated."

New generation of Kernel Visual Distinguishability (KVD)

As many are familiar, the basis for Canada's quality assurance system for wheat and other key grains is identification by Kernel Visual Distinguishability (KVD). In the case of wheat - where KVD is most prominent - this system requires that each wheat class have a different kernel shape or colour, readily identifiable by the human eye of grain inspectors. KVD has allowed for effective handling of wheat from farm to port for many years, but many in the grain industry feel the sophistication of this system needs a major upgrade to respond to the increased fragmentation of both products and markets.

This need has also intensified at a broader level. The Canada Grain Act and Regulations covers 21 different types of grain, all of which are segregated into three or more grades, each with unique end-use processing qualities. Markets are becoming more quality and variety specific every year. The prospect of genetically modified organisms (GMOs), innovative new quality types, and an overall boom in grain diversity add further pressures.

This is where technology such as Acurum is expected to provide dramatic benefits, says Coulter, heralding a new era of visual analysis to lead Canada's grain handling system into the future.

The technology behind Acurum was developed by AAFC lead researchers Eric Kokko, image analyst, and Dr. Bernie Hill, neural network modeler, along with AAFC crop pathologist Dr. Bob Conner. Researchers throughout the AAFC network and the Canadian Grain Commission also contributed to the development by providing grain samples and supportive expertise. On the commercialization end, Coulter worked along with Dr. Bert Innes of AAFC and Stan Benda of the Department of Justice. Here, Coulter reflects on Acurum's development, the innovations it features and what this type of new technology means for the future of Canada's grain industry.

A:The core technology is what we're calling a Visual Particle Analysis System (VPAS), which has many potential applications.

This technology was initially developed to solve research problems in the plant breeding field. Once these research problems were resolved, there was recognition that the technology held tremendous potential for similar applications in agriculture and other industries. Most immediately, this potential became apparent for seed analysis in the grain industry.

As a result, the first application of VPAS is a seed-analysis version called Acurum. Acurum is an automated system for grain analysis. It rapidly measures physical characteristics of grain kernels, such as texture, length, shape, width and colour characteristics. This information is then processed in a computer using artificial intelligence and neural network modeling.

Neural networks are computer software programs patterned the way medical researchers believe the human brain works. Researchers "train" the networks by inputting as many relevant examples of a given situation and as many variables as possible. The computer uses this data as experience, and when presented with a new situation, sifts through the data, mixing and matching the variables until it determines the most likely solution. The system can rapidly analyze samples of over 10,000 seeds at a time in two minutes with pinpoint accuracy.

A: Canada's long-time quality assurance system for cereal grains is based on human visual inspection. This is subjective, labour intensive and carries an inherent risk of inaccuracy and inconsistency. Acurum is an automated system that provides uniform, objective analysis in a much faster and far more sophisticated manner. The neural networks aspect of the technology allows it to perform more precise judgment-based analysis than other machine-vision technologies.

Q: Acurum has potential for wide-reaching impact. Where do you see the key applications and benefits?

A: As a major grain-trading nation, Canada stands to benefit enormously from this technology. It will dramatically boost our ability to deal with increasing pressure on the food-supply chain and place Canada at the leading edge of providing grain-quality assurance to its customers. This will benefit all players in the grain industry, from farmers to marketers, and strengthen Canada's position in the multi-billion dollar grain trade.

Most important, this technology will support a safe and sustainable Canadian food industry. An example of this impact is Acurum's ability to allow for precise detection of diseases in wheat, such as Fusarium Head Blight, that have the potential to make grain unfit for livestock and human consumption.

In my personal opinion, Canada has enjoyed a strong history of producing high-quality grains. From a practical perspective, Acurum will do two key things to support this industry. First, it will allow Canada to more accurately quantify some of the positive traits associated with our grain industry, through traceability of grain products.

Second, it will allow the grain companies to much better manage what product they have, by keeping precise track of what grain is in what silo, in which terminal, in which province, etc. The companies will know exactly what they've got and where it is. They can then use this information to mix batches A, B and C to come up with the most cost-effective package to market. We see Acurum proving a very powerful management tool from the grain companies' perspective.

A: Acurum as a commercial technology is now in the later stages of development. We expect some instruments to be delivered shortly to selected industry representatives for testing, and sometime after that we expect to have the commercial instruments available. The exact time of introduction is up to DuPont Canada, to which we've licensed the VPAS technology.

Ultimately, Acurum will be available across the grain supply chain, for potential use by farmers, elevators, terminals, Canadian Grain Commission, grain users such as flour mills, malting companies and foreign grain customers.

AAFC owns and has filed patents on the technology. DuPont Canada is the sole licensee. As part of the licensing agreement, the technology will be made widely available to Canadian industry and AAFC will receive royalties based on sales of the technology.

DuPont has been an important partner in this process. The development of VPAS for Acurum is a large project in terms of both time and cost. It's therefore essential to have a commercial partner with the wherewithal to invest in all the added costs of getting the technology into a marketable form and delivering it effectively to that market. DuPont has also brought to the table complementary technologies that will allow it to expand on VPAS in important ways that might not be possible for others.

Over the longer-term, there are many potential applications for the VPAS technology behind Acurum. The extent of those applications will be limited only by one's imagination. As a platform technology, VPAS has tremendous potential for applications on a very broad basis. Initially, the focus is expected to be on agricultural applications, followed by biological applications in other areas. Eventually, we expect there will be many industrial applications.

Even for Acurum itself, the current progress is just a starting point. There will be ongoing research and development efforts to enhance and deliver an improved version of the technology in the years ahead.

Reprintable with permission. Reproduction of this article - in whole or in part, in print or electronic - requires direct permission from Meristem Information Resources, Ltd. Contact Meristem directly to request reprint permission.

Archives

Inside the future of grain analysis

New generation of Kernel Visual Distinguishability (KVD)