Quebec municipalities use AI to track tree cover, cars, swimming pools

MONTREAL — Municipalities in Quebec’s metropolitan region have recently begun using artificial intelligence to track everything from tree cover to cars and even backyard swimming pools.

The Communauté métropolitaine de Québec, which groups Quebec City and its suburbs, says the groundbreaking project will help member municipalities meet environmental goals, assess parking availability and monitor urban development. But as more Canadian cities embrace AI tools, experts warn they will have to think carefully about how they use the technology and whether the public fully supports it.

Frédérick Lafrance, geomatics development manager at CMQuébec, said the organization trained a deep learning model on high-definition aerial images of Quebec City and the surrounding region taken in 2021. The AI ​​model can pick out and highlight a variety of features, including buildings, trees, vehicles, swimming pools, backyard trampolines and monkey bars.

“The performance is equivalent or very similar to a human,” he said in a recent interview. “Except the execution speed is much faster, so we can do a lot of work in a very short time.”

The data can be used in a variety of ways. Lafrance said Quebec City and many surrounding municipalities have goals for urban greening and tree cover, and artificial intelligence is a natural tool to measure progress. Conversely, it can also measure how much green space has been converted to asphalt over time.

Municipalities could also use the footage to see if there is enough parking in different areas, he said. And keeping track of backyard pools could help cities decide where to send inspectors.

“That information is hard to come by because it requires a lot of manual labor,” he said. “So we provide a service that shows them where the pools are, and they can use it to follow up.”

A new set of aerial images will be taken this summer, Lafrance said, and he hopes to analyze them this winter. He also hopes to apply the technology to photos taken as far back as the mid-20th century to assess how the region has evolved over time. The 2021 images are currently available through the provincial government’s open data portal.

Lafrance said that he believes CMQuébec is the first municipal agency in Quebec to use AI in this way. However, he believes it is only a matter of time before other major cities adopt similar tools.

Renee Sieber, an associate professor of geography at McGill University who studies artificial intelligence, says municipalities across the country are already using AI in different ways.

Edmonton, for example, has used artificial intelligence as part of a project that uses remote cameras to monitor wildlife entering the city. Montreal and Toronto have experimented with AI to ease traffic congestion, and Montreal’s public transit agency is piloting AI to prevent suicide on the city’s subway system by scanning CCTV footage for signs that someone is in distress.

Still, Sieber warned that some applications of AI raise more concerns than others.

“There’s a big difference between a tree and a swimming pool in the backyard,” she said, adding that cities could use such technology to catch minor violations — unfenced swimming pools, for example, or illegal sheds in people’s backyards.

“I’m always wary of cities that don’t have clearly defined objectives … because that increases the potential for mission creep,” she said. “Frankly, people don’t like being watched from above, and it’s very different from a building inspector walking around and peeking into someone’s backyard …. It creeps people out.”

In the case of Quebec City, Lafrance pointed out, the aerial images already exist, and the AI ​​isn’t revealing any information that couldn’t already be tracked manually—it would just be much slower. The image resolution is sharp enough to distinguish between, say, a car, an SUV, and a truck, but not to identify the make and model of individual cars, let alone license plate numbers.

There are many opportunities for municipalities that want to apply artificial intelligence to images. Google has used a combination of aerial imagery and AI to estimate tree cover in hundreds of cities around the world, including Toronto, Montreal and Vancouver.

Richard Khoury, a computer science professor at Université Laval, said the Quebec data could be used to assess the value of different properties and to select less developed areas for urban development projects.

But AI can reveal more than just the number of trees and cars in a neighborhood.

A 2017 U.S. study used deep learning to identify the make and model of cars in 50 million Google Street View images. The researchers found that cities where sedans outnumbered pickup trucks had an 88 percent chance of voting Democratic, while cities with more pickup trucks than sedans had an 82 percent chance of voting Republican.

The authors presented their findings as a powerful tool for researchers and policymakers, but also raised important ethical questions. “It is clear that public data should not be used to undermine reasonable privacy expectations of individual citizens, and this will be a central concern going forward,” they wrote.

Sieber said municipalities need to think about how the public will react to their use of artificial intelligence. “No tool is neutral,” she said. “When you talk about reliability in AI, there’s a question of performance … but there’s also a question of social acceptance. If you’re surveilling people invisibly, you’re not likely to increase social acceptance.”

This report by The Canadian Press was first published Sunday, July 28, 2024.

Maura Forrest, The Canadian Press