Pearson Airport tries out runway inspections powered by drones and AI

There was something unusual in the skies above the runways at Pearson International Airport this week — not a plane, but a drone.

It’s a trial run of a new technology designed to use AI to speed up runway inspections with an eye to allowing Canada’s busiest airport to get even busier.

• Download our app to get local alerts to your device
• Get the latest local updates right to your inbox

For about five hours on Monday, a crew with Victoria, B.C.-based Niricson Software huddled over a laptop tracking the drone as its camera took high resolution photos of the asphalt.

“We’re looking for cracking and spalling and anything that would give a sense of the integrity of the surface,” said Dean Arscott, the company’s UAV operations lead, in an interview off runway 15R/33L.

The camera’s images are then fed into a machine-learning algorithm that can identify any defects and build them into a detailed 3D model of the runway.

The advantage, said Arscott, is that crews manually inspecting it take longer and don’t build any digital records.

“Instead of taking the entire day to do this runway, we can do this in a few hours, and that means less time for the runway, more money for the airport,” Arscott said.

Runway condition is a serious issue for airports, who must keep their runways smooth as any stray objects can damage airplanes when they take off or land.

“This is a trial to see if this is something that could be the future of how we do these inspections,” said Tori Gass, a Greater Toronto Airports Authority spokesperson.

“This could really revolutionize the way we do inspections,” she said.

The project is funded by Transport Canada, said Arscott. The technology was tested out on an unused airport near Killaloe, Ont. and has also examined other infrastructure in B.C. including hydro dams.

The 3D model of an unused airport near Killaloe, Ontario created by Niricson Software (supplied).

“On one dam we identified over a million cracks. This would have taken a human working every hour of the day for two years straight to do themselves,” said Arscott.

If it works, he said he hopes the technology takes off.

“We could do every airport in the world,” he said. “We have big vision at our company.”