ONTARIO GRAIN FARMER MARKET DEVELOPMENT 30 The new varieties will be specifically bred by NRGene Canada and Pulse Genetics, not just for better adaptation to a shorter season or harsher growing environment, but also for improved yield and resistance to soybean cyst nematode. The second phase of the project will be the development of a novel, soy-based plant protein product. It’s the first single-ingredient plant-based protein of its kind, involving the extrusion of whole non-GMO soybeans. AUTOMATING DETERMINATION OF SOYBEAN QUALITY Also in June, PIC announced a partnership with Grain Discovery, Inarix, and Sevita International to produce a tool that uses artificial intelligence (AI) to instantly analyze soybeans. PIC has provided $700,000, and the partners will invest the rest of the $1.3 million needed in the project. Inarix will develop the platform, Grain Discovery will lead the commercialization and market deployment, and Sevita International will contribute data and serve as the primary site for real-world validation. This tool is basically an app. Users take a photo of a soybean sample using a standard smartphone, and in under 20 seconds, the variety is identified. “This is going to be valuable in the hands of the farmers as well as at the elevator,” explains Sandy Hart, general manager at Sevita. “At the elevator, it’s a fast and reliable way to confirm variety identification upon delivery. In identity-preserved (IP) crops, this can be a very valuable ‘last line of defence’ that really doesn't exist today beyond a visual inspection when it comes to confirming the variety on the scale.” That is, Hart acknowledges that farmers themselves always know what they are growing, but deliveries can often be handled by farm employees with varying levels of experience and/ or third-party trucking companies. He notes that it’s rare for a load of the wrong variety to be unloaded into the wrong bin, “but it does happen, and it’s very costly when it does.” For farmers, Hart notes that the most common quality challenge in IP crops is poor aesthetics (dirt tag, for example) caused by harvesting in marginal conditions and/or heavy weed pressure. So, “in addition to the variety identification piece, we expect the tool will be able to give farmers reliable, real-time feedback on the visual aspects of what they are harvesting,” he says. “This should be very helpful to them in terms of maximizing the window they have to work with for harvest while ensuring they are meeting important quality requirements.” Rory O’Sullivan, commercial leader for agriculture at Grain Discovery, describes the tool as another step forward in making the grain supply chain more transparent and efficient, and wishes that it had been around 20 years ago when he was out sampling grain. He notes that over 15,000 images of soybeans were used to ‘train’ the AI to discern differences in soybean shape, size, and microscopic details that all help in highly reliable variety identification. He predicts that in future, the model will be expanded to other visual parameters related to quality, such as colour, staining, and presence of foreign seeds. ““Canada can, and in a lot of ways already does, lead the world in producing high-quality soybeans and many other important grains to be used directly in making food.” continued from page 29
RkJQdWJsaXNoZXIy MTQzODE4