An undergraduate student in India is working with a New Brunswick university to help Canadian farmers manage weeds effectively and maximize crop yields using artificial intelligence (AI) technology.

Hina Tomar, an electrical engineering researcher at Aligarh Muslim University in Aligarh, India, is working remotely with Université de Moncton engineering professor, Habib Hamam to develop AI-equipped drones that can detect “even the smallest” of weeds, Hamam explained in a news release.

“The software functions as the brain of the drone, so adding the AI component is an important first step in advancing this next wave in farming technology,” said Hamam, adding that Tomar is currently working on accessing databases containing thousands of images of weeds to train the system to identify different varieties and sizes.

 

 

The technology will analyze images of crops in a field captured by a drone, filtering dirt, shadows and other impurities to distinguish between vegetation and weeds. The location of the weeds can then be sent back to the drone which applies herbicide only where it’s needed.

Researchers have been using drones to identify weeds, Hamam said, but the current approach typically only detects large, dense patches rather than individual weeds, adding that Tomar’s software will be more precise and efficient than others, which rely only on image processing.

Tomar’s research is part of the Mitacs Globalink program, a federally and provincially funded initiative, which pairs international students and researchers with Canadian research institutions. Tomar is one of over 1,000 students from 12 countries taking part in the program this summer across a range of industries.

Although she wasn’t able to travel to Canada due to the pandemic, Tomar meets regularly with Hamam through Google Meets and hopes to have a prototype ready by the end of her 12-week internship in August, at which point testing will begin at local Moncton farms. •