The City of Thunder Bay is planning to modernize how it tracks potholes by turning to artificial intelligence.
The city has issued a request for expressions of interest for an automatic pothole detection system, aiming to make road maintenance more efficient. According to Ian Spoljarich, the city’s roads manager, the new technology would replace Thunder Bay’s current manual inspection process, which still relies on pen, paper and logbooks.
“The system we’re looking to implement uses a smart camera mounted on city vehicles to automatically scan and analyze the road conditions as the vehicle drives, and records any of those deficiencies and sends them back so that we can review afterwards,” said Spoljarich.
The goal, Spoljarich said, is to improve how the city tracks road deterioration over time.
“We don’t see huge efficiencies for repairs, but more of tracking and logging what deficiencies they are,” he said. “This will catalogue any potholes, cracks, infrastructure deficiencies that we see on the roadways, and it’ll actually take a picture of that, which we haven’t been doing before.”
Ian Spoljarich, manager of roads with the City of Thunder Bay, says AI technology could help the city better track road deterioration. (City of Thunder Bay)
According to the city’s request document, Thunder Bay is seeking a camera or sensor-based system that can automatically detect potholes, geotag them, generate repair work orders and integrate with the city’s maintenance management software. Privacy safeguards are also a key requirement. The report states the system must automatically discard unnecessary data, such as license plates numbers or people’s faces.
The automated system would be installed on two vehicles currently assigned to road patrol. If a vendor is selected soon, the city hopes to have the equipment in place before winter.
“We want to get four-season coverage just to start using the software,” Spoljarich said.
The project is being funded through the city’s existing capital and operating budgets. Details on cost have not been released because the contract has not yet been awarded.
Thunder Bay is following in the footsteps of other municipalities that have already adopted similar systems.
A common tool for other Ontario cities
Durham Region, Markham, Kingston, Vaughan and Richmond Hill have all piloted CityROVER AI. It’s a smartphone-based system that automatically identifies potholes using vehicle-mounted cameras and sensors.
According to Khalid Elgazzar, Canada Research Chair in the Internet of Things at Ontario Tech University, there are three approaches. Camera-based systems use mounted cameras to capture continuous road images, sensor-based systems use vibration data from vehicles, and smartphone-based systems crowdsource data from drivers. Each can automatically map potholes and track deterioration.
Khalid Elgazzar is an associate professor and Canada Research Chair in the Internet of Things with the Faculty of Engineering and Applied Science at Ontario Tech University. (Ontario Tech University )
“Camera-based systems may suffer from poor lighting, bad weather conditions, or sensor noise, which can cause false detections or missed potholes. Some system designers also use a hybrid approach that may combine some of the above to improve the accuracy,” he explained in an email to CBC.
He said the benefits for cities and residents are significant.
“These systems enable faster, data-driven road maintenance,” Elgazzar states. “This leads to quicker repairs, safer and smoother roads, fewer vehicle damages and improved public satisfaction. For cities, it supports better planning and resource allocation.”
He added that concerns around privacy are valid but largely manageable.
“Some residents also worry about data storage and access, including how long images or GPS data that may contain their information are kept and who can view them,” he said. “Camera-based or smartphone-based methods can capture images near homes or vehicles, though most systems blur faces and license plates.”
