SASKATCHEWAN POLYTECHNIC
Ground Truth Agriculture and Saskatchewan Polytechnic collaborate to advance AI systems for grain quality assessment

Ground Truth Agriculture Inc., a pioneering agtech company specializing in real-time grain quality data, has joined forces with Saskatchewan Polytechnic's Digital Integration Centre of Excellence (DICE) to fine tune their cutting-edge artificial intelligence systems. The collaboration aims to advance their innovative, automated grain grading system, to provide farmers and buyers with accurate and reliable grain quality information.

Image Credit: Ground Truth Agriculture Inc

"Sask Polytech was the first place we contacted for support with artificial intelligence and machine learning," says Divyesh Patel, Ground Truth Agriculture chief operating officer. "The applied research work Sask Polytech has completed in the past five to seven years is impressive. Working with the DICE team has been seamless and having access to their researchers and expertise has meant we have been able to develop an award-winning system faster."

Traditionally, the grading process for determining grain quality has relied on assumptions made from small samples, leaving producers with an incomplete understanding of their overall crop. Ground Truth Agriculture developed a machine learning platform that can improve the speed and accuracy of grain quality assessment at harvest and all points of the grain supply chain. This automated system benefits farmers by providing them with more accurate pricing information, while ensuring buyers have a comprehensive understanding of the grain quality they are purchasing.

The collaboration between Ground Truth Agriculture and DICE has already yielded impressive results. "The first project we completed with DICE was on Canadian Western Red Spring wheat, using machine learning to identify 50 visual characteristics of this crop," shares Patel. "The DICE team helped build a model that detects the visual characteristics and freed Ground Truth Agriculture employees to focus on other important tasks." The two DICE interns who worked on the Canadian Western Red Spring wheat project were funded by Mitacs.

Building on this success, with the help of a Natural Sciences and Engineering Research Council of Canada (NSERC) Applied Research and Development grant and Innovation Saskatchewan's Agtech Growth Fund, recent Sask Polytech Computer Science Technology (CST) graduates are now working to develop a similar detection model for red lentils. By utilizing machine learning and machine vision technologies, the DICE researchers are helping create a system to grade red lentils by visually identifying diseased and damaged kernels and impurities in the sample.

"The DICE researchers complete a lot of the groundwork, which can be very time consuming," says Patel. "It takes a lot of time and effort to figure out what is and isn't working. It's good experience for a junior machine learning engineer interested in the agtech industry. The DICE team has also helped us secure grant funding to continue this research, which is incredibly beneficial."

"Partnering with Ground Truth Agriculture offers Sask Polytech students and recent graduates remarkable internship opportunities," says Tanya Lung, DICE research manager. "DICE interns gain exposure to new technologies and acquire insights into applying their skills within specialized industries like agtech. Collaborating with Divyesh and his team, our interns delved deep into the agtech sector, exploring aspects such as grain grading, different grain types and harvest time. They also collaborated with the Ground Truth Agriculture team in leveraging various computer vision tools and methodologies."

Prakher Singh, DICE researcher for Ground Truth Agriculture, shares, "It's exciting, it's not often that grads get the chance to delve into machine learning right after graduation. I'm grateful for the offer from DICE and Ground Truth Agriculture. Each day, I'm learning and collaborating with a highly innovative team."

Ground Truth Agriculture envisions its systems being used in grading labs, on farms, at elevators, testing organizations, and even during international sales on boats and trains. The versatility and reliability of this agtech system make it an invaluable tool for all stakeholders in the grain industry.

"Sask Polytech's intellectual property policies also make it an attractive partner," adds Patel. "Intellectual property for applied research projects is retained by industry, meaning that Ground Truth Agriculture maintains full control over the IP on our joint research projects. This is a great comfort."

Once the visualization work is completed on red lentils, Ground Truth Agriculture and Sask Polytech look forward to collaborating on more models to visualize the next grain together.

DICE is Saskatchewan's first Technology Access Centre (TAC) funded by NSERC and Innovation Saskatchewan.  DICE performs applied research in a diversity of sectors focused on the manipulation, analysis, storage, transmission or display of data. Data is at the core of what DICE does. Learn more at saskpolytech.ca/dice.