A new neural network developed by NtechLab (a partner of Rostec) is capable of detecting pits, potholes, missing manholes, and damaged barrier fences on roads.
The system automatically notifies the responsible authorities about the identified problems. According to the developers' idea, this will allow for the prompt elimination of defects. The innovation is aimed at reducing the number of road accidents and minimizing damage to vehicles. The algorithm is capable of detecting accumulations of garbage near roads, overflowing containers at bus stops, and damage to buildings.
The General Director of NtechLab, Alexey Palamarchuk, noted that this development will make the urban environment more comfortable and safe. In his opinion, timely detection of road surface defects will help to significantly reduce the number of accidents.
The neural network analyzes video streams from cameras installed on public transport. This makes it possible to accurately identify and eliminate problems in any part of the city. The AI records non-working streetlights and damaged road signs. A pilot project is planned to be launched in one of the Russian regions in the coming month.
Read more on the topic:
Ntechlab's neural network taught to recognize people after decades
AI takes to the roads: AI will wean drivers off typing messages while driving
Cameras to improve traffic on roads launched in the Sverdlovsk region