Neural network will control the cleaning of icicles and snow in St. Petersburg

Artificial intelligence will detect uncollected icicles and snow on the roofs of non-residential buildings in St. Petersburg in winter. This was reported by the press service of the State Administrative and Technical Inspection on Thursday, October 2.

At the end of the 2024-2025 winter season, GATI employees completed the training of eight Gorodovoy neural network complexes. The task of the system is to accurately distinguish ice and snow from stucco and bas—reliefs on the facades of buildings. When the accuracy became high, the function was put into operation.

As he writes 78.ru In 2024, 21 owners of non-residential premises received fines totaling 2.1 million for late cleaning. In the coming winter, the neural network will start checking with the first snow, especially where violations have already been detected.

Earlier it was reported that in St. Petersburg, AI was taught to automatically determine the growth sites of Sosnovsky hogweed. If a dangerous plant is detected, the owners are fined 100,000 rubles.

All important news is on the Izvestia channel in the MAX messenger.