By Janan Shoja Doost
A researcher at the University of Guelph is utilizing artificial intelligence (AI) and hyperspectral imaging to transform soybean breeding, enabling faster, smarter and more precise crop selection.
Dr. Milad Eskandari, a professor in the Department of Plant Agriculture in the Ontario Agricultural College, leads a breeding program that combines traditional breeding methods with advanced AI and remote sensing technologies to develop high-yielding, disease-resistant soybean cultivars for both Ontario and international markets.
In a recent study published in Remote Sensing, Eskandari and his team used AI and hyperspectral imaging to predict soybean yield and biomass at early growth stages.
“I consider AI a powerful tool,” says Eskandari. “It allows us to better analyze the complexity of plant genetics and environmental stress, especially in cases where classical methods fall short.”
The study demonstrated that hyperspectral vegetation indices combined with hybrid AI models – including deep neural networks – can accurately predict soybean yield and biomass. These technologies enable breeders to detect early-season stress and reduce the need for time-consuming and labour-intensive manual inspections of several thousand soybean lines.
These advanced tools are already reshaping Eskandari’s breeding program.
Each year, he selects around 2,000 promising lines from a pool of over 15,000. With drone imaging and machine learning, that process is now significantly faster and more targeted. Eskandari’s lab is currently developing image-based AI tools to classify key traits, such as hilum colour, for food-grade soybeans – a project led by PhD student Katherine Fortune.
For farmers, digital tools such as AI, hyperspectral imaging and remote sensing provide powerful insights. Remote sensing can identify problems like soybean cyst nematode (SCN), which can significantly reduce yield, even before symptoms appear. Eskandari says these tools can help farmers monitor fields more efficiently while helping breeders select more resilient varieties.
“We’re not trying to promote AI for the sake of it,” he says. “We’re using it because it helps us tackle complex biological questions and make better, faster decisions. That’s where its real value lies.”
The project is funded by Grain Farmers of Ontario (GFO), the Natural Sciences and Engineering Research Council (NSERC), Mitacs, SeCan and the Ontario Agri-Food Innovation Alliance, a collaboration between the Government of Ontario and the University of Guelph. This research takes place, in part, at the Ontario Crops Research Centre – Ridgetown, which is owned by the Government of Ontario, through its agency Agricultural Research and Innovation Ontario, and managed by U of G through the Alliance.