Neural network taught to identify hogweed in St. Petersburg

In St. Petersburg, a neural network has been taught to automatically identify the growing sites of a dangerous invasive plant, the Sosnovsky hogweed. This was announced on Tuesday, August 12, according to the press service of the GATI.

The inspection specialists showed the artificial intelligence all the stages of weed growth so that it could learn to find it at the earliest stages. As soon as a high percentage of detections was recorded, the function was put into operation.

Now eight mobile neural network complexes "Gorodovoy" are operating daily in the Northern capital. As he writes 78.ru They monitor the territory along 186 routes using a new feature.

If the hogweed is detected automatically, the owners are fined 100 thousand rubles. It must be paid for and the weed removed. The neural network will check the fulfillment of the requirement during the second detour.

At the end of July, Russian President Vladimir Putin approved a law that obliges land owners to combat Sosnovsky hogweed and other alien plants. The new rules will come into force on March 1, 2026.