University of Waterloo researchers develop new system to detect pothole problems
Local researchers at the University of Waterloo have developed a new artificial intelligence system to detect problems with roads.
Currently it can take up to 24 people to go through about 13,000 images a day looking for potholes. But the new research suggests computers can do majority of the work faster, and with more accuracy.
All the computer needs is images of the roads that can be taken using something as simple as a smart phone, and then their algorithm will do all the work.
“All you have to do is put it through our algorithm and we can output the depth of the crack, the extent of the crack, the severity of the defect,” said PhD candidate David Abou Chacra.
The City of Waterloo typically budgets around $200,000 every year for pothole repairs, money that experts with the Vision Image processing Lab at the university say could be saved using the new technology.
But the researchers are hoping to take it one step further than just road repairs, wanting to use the technology to assess bridges, buildings and other assets for municipalities.
“This is actually an alternative to GPS measurements where it is much more accurate and you can get way more accuracy in this kind of situation,” said PhD candidates George Younes.
The team is currently performing field trials to see how well their system works.
They will also be teaming up with the Ministry of Transportation for a yearlong project to see how it works on images of provincial highways.