A new app, developed at the University of Guelph’s Ontario Agricultural College, aims to help wheat farmers across the world. The free mobile app is called WheatScanR. Farmers can simply take a picture of their crop and receive near-instant data on one of the most common and destructive diseases affecting wheat production, Fusarium head blight (FHB). “Basically, what it does is it uses deep learning and AI to find every single wheat head within a cellphone image and count them,” said Riley McConachie, a PhD candidate in plant agriculture at the University of Guelph. The idea came out of a desire to free farmers from the labour-intensive research work by harnessing the power of artificial intelligence. “We focus a lot on [seed] breeding for resistance to the disease, and one of the ways we have to do that is evaluating plots out in the field and testing if the lines are resistant to the pathogen or not,” McConachie explained. “This is really time consuming with often thousands of plots, and it’s very subjective. You just kind of guesstimate how much infection is in the plot. We’re trying to develop a way that’s a lot more accurate at determining the degree of infection within each plot.” McConachie said any farmer with a smartphone can download the app. All they have to do is take a picture and they will get near-instant data about the health of their crop. “Farmers can use this to help gauge how much infection is in their field, and it allows them a little bit of leeway so that they can employ some different practices to minimize the impact of the disease,” McConachie said. The data is also kept private. “All the user’s data is their own, it’s completely 100 per cent their own. We don’t use any of it.” The app developer said keeping WheatScanR free and accessible is in line with the open-source model that was essential to their work. “We benefited a lot in our research from open-source models, and we’ve taken advantage of this in order to actually employ all of this. We just wanted to continue that trend and give back.” McConachie said that, since they launched last August, the app has already been downloaded 250 times. An improved version of WheatScanR should be ready ahead of the 2027 season.
