Pioneering AI-powered waste audits

Doug MacMillan
Doug MacMillan • 2 February 2022
guelph garbage truck

How much avoidable food waste and non-organic material are Guelph residents tossing in their green cart? In October 2020, the City of Guelph launched a Residential Waste Data Challenge to find out. Eagle Vision Systems responded.

The Kitchener company had worked with the City in the past to automate other aspects of waste collection by developing CartSeeker, an automated arm operation. Leveraging similar technology, this time they set out to develop a first-of-its-kind system to analyze organic waste in real time as each household green cart is emptied into the collection vehicle.

Working with researchers at the University of Guelph’s Intelligent Control and Estimation (ICE) Lab, Eagle Vision developed a video system that records material going into the trucks. From there, Artificial Intelligence algorithms were used to detect five target items: compostable bags, non-compostable bags, yard waste, recyclables, and avoidable food waste.

Two summer students reviewed thousands of video stills to train the AI to identify target items. By the end of the pilot project, the technology could identify these items with 90 per cent accuracy. “It is amazing the technology that is coming from these projects,” says Chad Scott, Manager of Logistics and Site Operations at the City of Guelph Solid Waste Resources Division, “We are only beginning to learn how to leverage this amazing work to improve safety, help our environment by reducing contaminants, and help our team raise the bar on service.”

This high-tech initiative provides a lot more knowledge about exactly what is being thrown away and in which neighbourhoods, without the need for manual waste audits. As a result, it can help inform public education programs and interventions aimed at reducing food waste in the green cart. And, because Guelph’s green carts have RFID tags that link them to their street address, City staff could also provide targeted interventions to areas of the city producing the most unnecessary food waste.