Quebec municipalities use artificial intelligence to track trees, cars, swimming pools

Article Contents

A municipality in Quebec's capital region recently began using artificial intelligence to track everything from trees to cars and even backyard pools.

The Consejo Metropolitano de Quebec, which represents 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 Canadian cities adopt AI tools, experts warn that they need to think carefully about how they're using the technology and whether citizens are fully on board.

Article Contents

Frédéric Lafrance, geographic information development manager at CMQuébec, said the organization trained a deep learning model on high-resolution aerial imagery of Quebec City and the surrounding area taken in 2021. The AI ​​model can extract and highlight a variety of features, including buildings, trees, vehicles, swimming pools, backyard trampolines and jungle gyms.

“Their performance is comparable to or very similar to humans,” he said in a recent interview. “They can execute much faster, so they can get a lot of work done in a very short amount of time.”

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

He said local governments could also use the imagery to see whether there are enough parking spaces in different areas, and tracking backyard pools could help cities decide where to send inspectors.

“This information is hard to get because it requires a lot of manual work,” he said, “so we're providing a service that shows you where the pools are so you can follow up.”

Article Contents

Article Contents

LaFrance hopes to capture new aerial images this summer and analyze them this winter. He also hopes to apply the technique to photos taken in the mid-20th century to evaluate how the area has developed over time. The 2021 imagery is now publicly available on the state government's open data portal.

Lafrance believes CMQuébec is the first municipality in Quebec to use AI in this way, but thinks it's only a matter of time before other major cities adopt similar tools.

Renee Seeber, an associate professor of geography at McGill University who researches artificial intelligence, said local governments across the country are already using AI in a variety of ways.

Edmonton, for example, is using artificial intelligence as part of a project that uses remote cameras to monitor wildlife entering the city, Montreal and Toronto are experimenting with AI to ease traffic congestion, and Montreal's transit agency is working on a pilot project to use AI to prevent suicides in the city's subway system by scanning security camera footage for signs that someone may be in distress.

Article Contents

Still, Seeber warned that some applications of AI are more concerning than others.

“There's a big difference between a tree and a backyard pool,” she said, adding that the city could use such technology to catch minor violations, such as a pool without a fence or an illegal shed in a backyard.

“I'm always wary of cities that don't have a clear mission because those cities are more likely to have an expanding mission,” she said. “Frankly, people don't like being watched from above. That's very different than having a building inspector walking around and peering into someone's backyard. … It scares people.”

In Quebec's case, Lafrance noted that the aerial imagery already exists, and the AI ​​isn't revealing information that can't be tracked manually — it's just that manual tracking would be much slower. The image resolution is clear enough to distinguish between cars, SUVs, and trucks, for example, but it can't identify the make and model of individual vehicles, much less license plate numbers.

Applying artificial intelligence to imagery opens up a wide range of possibilities for cities: Google has combined aerial imagery with AI to estimate tree cover in hundreds of cities around the world, including Toronto, Montreal and Vancouver.

Article Contents

Richard Coury, a computer science professor at Universite Laval, said Quebec's data could be used to assess the value of different properties and to target underdeveloped areas for urban development projects.

But AI has the potential to reveal much more than the number of trees and cars in a neighborhood.

A 2017 US study used deep learning to identify car makes and models from 50 million Google Street View images. The researchers found that cities with more sedans than pickup trucks had an 88% chance of voting Democratic, while cities with more pickup trucks than sedans had an 82% chance of voting Republican.

The authors presented their findings as a powerful tool for researchers and policymakers, but also pointed to important ethical issues: “Public data should clearly not be used to undermine individual citizens' reasonable expectations of privacy, and this will be a central concern going forward,” they wrote.

Seeber said local governments need to consider how the public will react to the use of artificial intelligence. “There are no neutral tools,” he said. “When we talk about the trustworthiness of AI, there's the issue of performance, but there's also the issue of social acceptance. If you're invisibly monitoring people, you're probably not going to have a high level of social acceptance.”

Article Contents

Article Contents