UNB researchers mapping traffic behaviour using satellite
Can track speeders 770km from Earth
Researchers at the University of New Brunswick have developed a technology that can analyze traffic behavior, from 770 kilometres above Earth.
By aligning photographs taken by a specialized satellite high in Earth's orbit, Yun Zhang and his team of researchers in UNB's geomatic department have developed algorithms that can detect moving objects, such as a vehicle, and calculate speed trajectory and direction.
"The satellite takes two images," said Zhang. "One in black and white, high-def, and one colour in low resolution. It then uses the small angle difference to find moving objects."
By averaging the speeds of all vehicles, Zhang and his team can build maps of speeding hot-spots, and areas of congestion.
This data can then be used by road builders to streamline traffic flow, or by police to focus speed traps.
The traffic maps can cover massive amounts of traffic quickly.
"The satellite can take images that each image is 16 kilometres by 16 kilometres," said Zhang. "From one shot you can see all the cars in that area.
"For example, Fredericton, we just need one image to cover the whole area. We see all the cars there."
Hoping to expand applications
The WorldView-2 spacecraft, launched in 2009, travels 7.1 km/second, circling the planet about once every 100 minutes. It snaps thousands of pictures as it goes.
The imagery is precise enough to detail traffic behaviour in individual highway lanes.
It can give data on the speeds of vehicles moving onto exit ramps, which was demonstrated Wednesday with the Oromocto exit off the Trans-Canada Highway and the fast lane parallel to that area.
The fastest vehicle tracked during the demonstration of the technology was travelling 162.6 km/h.
Individual licence plates cannot be seen from space due to their placement on the front and rear of vehicle bumpers. But if identifiers were placed on the roofs of vehicles, identifications could be made.
Zhang’s graduate student, Zhen Xiong, is working to take the technology even further.
"We can extract other information from the images, like the types of vehicles," said Xiong. "From types of vehicles, we can estimate the pollution they put in that environment."