Russia creates neural network to assess coastal pollution

Russian specialists have developed a neural network to analyse pollution on the shores of water bodies. This is reported by the official website of the Ministry of Natural Resources and Environment of Russia.

It is noted that the model has already been tested in the Kronotsky reserve in Kamchatka. There stressed that some places are difficult to access, which complicates the organisation of cleaning.

"It is important to understand what resources need to be involved. At first we launched drones to film the coast, to see the area of rubbish. Then we decided to create a neural network that will help us calculate the amount, volume, weight and classify rubbish by type so that coastal cleanup goes faster and more efficiently," said Roman Korchigin, the author of the project, an employee of the Kronotsky Reserve.

The technology analyses photos of the coastline, classifying rubbish into six types: fishing nets, metal, rubber, large plastics, concrete, and wood. The model's accuracy is above 80 per cent.

Photo: iStock