Russia Deploys AI to Detect and Remove Plastic Waste from Illegal Dump Sites

A new vision-based neural network developed in St. Petersburg can spot PET bottles with 92% accuracy—and pinpoint them on a digital map.

Researchers at Saint Petersburg State University of Aerospace Instrumentation have developed an AI system capable of identifying polyethylene terephthalate (PET) bottles at illegal dump sites. The technology uses video feeds from surveillance cameras and has achieved a 92% detection rate, according to Gazeta.ru.

The system doesn’t just recognize plastic waste—it also maps its exact location. The next step is a fully automated cleanup operation: one robot will scout the area and generate a map, while a second robot will travel to the coordinates and collect the waste. Future improvements will integrate additional sensors to boost detection accuracy and field performance.

AI has become a central pillar in Russia’s environmental strategy. Neural networks are already being used to detect wildfires and catch illegal dumpers, using images from cameras deployed in urban areas, nature reserves, and remote wilderness zones.