AI That Doesn’t Cross Lines—It Cleans Coastlines
A neural network that detects and classifies trash in aerial photos is helping volunteers map and clean polluted shorelines—turning artificial intelligence into a practical environmental tool.
Artificial intelligence is starting to play a hands-on role in Russia’s environmental cleanup efforts. Yandex reports that since 2024, AI tools have helped volunteers clean more than 50 kilometers of coastal territory. The work is part of a collaboration between the Clean Shore volunteer movement, Yandex specialists, and researchers at Far Eastern Federal University.
Finding Trash From the Sky
At the core of the project is a neural network that analyzes aerial photographs of shoreline zones. The algorithm identifies waste and sorts it by category—including fishing nets, metal, plastic, rubber, concrete, and wood. Its detection accuracy exceeds 80 percent. The results are automatically plotted onto Yandex Maps, displaying coordinates, estimated volume, and composition of debris. This helps volunteers assess cleanup scale, plan routes, and select the necessary equipment.
A Purpose-Built Neural Network
The system was developed by Yandex’s Center for Technologies for Society and the company’s School of Data Analysis, with scientific support from Far Eastern Federal University. In 2025, the model was retrained on 20,000 additional images to improve recognition accuracy and expand its analytical capabilities. According to project organizers, AI has shortened expedition-preparation time severalfold and made cleanup operations significantly more effective.
For Reserves and National Parks
In 2025, nature reserves and national parks joined the initiative. In 2026, the project’s participants plan to shift from isolated cleanup events to systematic, year-round coastal maintenance. Special focus will be placed on Arctic territories such as Novaya Zemlya and Franz Josef Land, where ocean currents bring large amounts of debris and cold conditions slow natural decomposition.
The project shows that AI can function not only as a data-analysis tool but as a real, on-the-ground assistant in solving environmental challenges.
