Tyumen Researchers Develop Neural Network to Detect Pipeline Gas Leaks
The system can operate around the clock by processing data from multiple monitoring sources.
Researchers at the University of Tyumen have developed an intelligent system that automatically detects gas leaks from pipelines. The software combines data from infrared cameras, gas analyzers, and wind sensors, allowing it to operate continuously without human intervention.
According to the developers, the new system could significantly improve infrastructure safety. It is designed to outperform conventional field inspection teams, which may not always detect small leaks or minor damage quickly enough.
The neural network was trained on combined datasets from sensors and video cameras. As a result, the system can now not only detect incidents but also distinguish between an actual emergency and a malfunction of monitoring equipment.
