Kazakhstan to use artificial intelligence to assess road conditions

Kazakhstan to use artificial intelligence to assess road conditions

In Kazakhstan, major repairs have begun on the Yekaterinburg-Almaty highway of national significance, covering a section more than 30 kilometers long. An analysis of the road surface conducted using artificial intelligence revealed that certain areas did not meet established standards. Specialists are now checking the strength and smoothness of the asphalt with the help of a mobile laboratory equipped with four cameras and smart sensors. According to experts, the new automated system has greatly simplified the monitoring process: previously, processing data for every 100 kilometers took four days, whereas now it takes only three to four hours.

«It was initially planned to carry out medium repairs, but after the diagnostics, the artificial intelligence system recommended major repairs. Accordingly, a decision was made to proceed with a full-scale overhaul,» noted Zholdybai Tentekov, Director of the branch of the National Center for Quality of Road Assets.

Defects in the road surface are detected by another invention created by Kazakh developers. For this purpose, Jolscan uses computer vision technology along with video footage from ambulance and patrol car dashcams. All the collected information is automatically compiled by the system into a damage map, which is then promptly sent to the city’s public utilities. To date, around 500 problem areas have been identified, more than half of which have already been repaired.

«We monitor roads using cameras installed on ambulances, specifically the ones facing the roadway. We collect their video feed, which is then processed on our server. A standard AI model detects potholes on the roads. The results are displayed on an interactive map with markers placed throughout the city. When you click on a marker, you can see an image of the pothole along with its geolocation, time, and date. Essentially, we’ve moved from manual data collection to an information-driven system,» shared project developer Rustem Kadyrov.