AI Road Management System Helps Detect Surface Conditions & Snow

With things like climate change, growing population, and increasing traffic, our roads and highways are under more and more pressure every year. Officials often can’t keep up with identifying and fixing deteriorating surface conditions fast enough. One company is using artificial intelligence to detect road conditions to help prevent accidents and further damage.

Chiba, Japan based Weathernews Inc. has recently succeeded in creating a prototype of an “AI road management support system”. The system can recognize and analyze road conditions in real time using video images. It then creates a 3-D model of the surface if further analysis is needed.

Through deep learning, the AI was able to recognize differences in image data detecting things such as snowfall, ice, and road damage. It can even detect conditions at night.

The AI road management system also factors in things such as camera type, angle, location, and brightness in making an analysis. Human eyes often have trouble taking these into account when assessing road conditions. For example, during testing, the AI was able to recognize even the most subtle differences in snow quality.

Morioka City Tests AI Road System

The company initially tested the system around Morioka City, Japan late last year. They used a dashboard camera which sent the footage in real time to be analyzed. It was able to detect early road damage and surfaces that could possibly cause an accident.

Last year the city had record snowfall. City officials discovered 4,800 cases of road surface damage which led to about 2000 points of road puncture (10 times the average amount) Had the city used Weather Inc.’s AI system to detect road damage earlier, it would have likely cut down on the number of punctures and accidents saving both lives and money. The AI road detection system is still in the prototype phase but the city plans more testing this winter when snowfall will be at its heaviest.