The Future of Food

New online portal makes machine learning in Canadian agriculture a reality

TerraByte, a research group from the University of Winnipeg, is launching an online portal later this month to share their data sets and pave the way for machine learning in Canadian agriculture. 

Led by Professor Christopher Henry and Professor Christopher Bidinosti, TerraByte started collecting data in 2020 using an imaging robot that captures and automatically labels images of crops found in the Canadian prairies from all angles, and records their metrics. They collected over two million plant images - using them to build machine learning models that can classify the type of plant, distinguishing between weeds and plants that look almost identical. 

The plant-image data and machine learning models will be accessible through the EMILI portal for academics and others in the industry to build and train their own autonomous machines. Through the portal, users can tailor their search down to the plant type, camera angle, image rotation, and camera position, before downloading it onto their system. 

Professor Henry told the CityAge event, Data to Drive Better Food Outcomes, that TerraByte is currently working on other models that will detect disease, and perform localization in the field as well as phenotyping. All of these models will eventually be accessible through the EMILI portal. The team, Professor Henry said is looking forward to seeing what other people do with the data to accelerate machine learning in the agriculture sector.

 Sign up for EMILI’s newsletter here to be notified when the portal goes live and get instructions on accessing the data.


Vancouver | Las Vegas

cityage newsletter: 
how to build the future

follow us