AI Detects Traffic Anomalies, Improves Flow and Reduces Environmental Impact

Russian Scientists Develop AI Program to Detect Traffic Anomalies

Researchers at South Ural State University (SUSU), in collaboration with TV BRICS, have developed a groundbreaking program that utilizes neural network technology to detect traffic anomalies. This program analyzes CCTV footage, accurately recognizing and tracking vehicles with a precision of up to 30 centimeters. This allows for the creation of real-time visual maps highlighting traffic flow difficulties.

The program's unique ability to detect even minor deviations in traffic, including lane width reductions, sets it apart. It effectively signals potential obstacles such as accidents or repairs, providing crucial information for proactive response. Intersection visualizations are updated every two seconds, with color-coded congestion indicators – the more cars in an area, the redder the box.

The program's capabilities extend beyond anomaly detection. In the future, the neural network will be able to classify anomalies by type, predicting potential congestion and its impact on traffic within the next 10-20 minutes. This valuable information will empower relevant departments and services to promptly address road issues and prevent further complications.

A significant advantage of this development is its easy integration into existing city infrastructure. Unlike other systems, it does not require the installation of expensive GPS sensors on each vehicle.

This project forms part of an intelligent transport monitoring system already operational in several major Russian cities. The implementation of this new program is expected to significantly improve urban traffic conditions by increasing traffic speed and reducing environmental impact.