Harnessing AI for Early Wildfire Detection in Canada 

University of Toronto ECE students use satellite imagery to develop an automated wildfire detection system for rapid, life-saving alerts

Climate change and global warming have led to an increase in wildfires across Canada, posing serious environmental and financial risks. While 2023 was a record-breaking wildfire season, according to the CBC, the 2024 wildfire season was on track to be the second-worst wildfire season in terms of area burned since 1995, with more than 5.3 million hectares burned so far.  2024 saw a concentration of fires in Western Canada; wildfires in Jasper resulted in $880 million in insured damages. 

Forrest Zhang, Yan Zhang and Eric Zhao are three master’s students from the University of Toronto’s Electrical and Computer Engineering department who have developed a project to help detect wildfires using satellite imagery. 

The primary goal of their project is to develop a system that can automatically detect wildfires in their early stages through satellite images, enabling rapid response without requiring human intervention.  

Since satellites capture constant images of forested areas, the team’s solution leverages this data to identify burned regions and segment them accurately. Early detection through this automated system could significantly reduce financial losses, mitigate environmental damage, and potentially save lives by providing timely alerts before wildfires escalate. 

This project is just one of the ways that ECE students are working to create a more sustainable future through technological advancements.