Abstract :

Natural gas leakage poses serious risks including fire hazards, explosions, and environmental pollution. Traditional gas detection techniques often fail to provide early and accurate leakage localization, especially in large-scale industrial environments. This project proposes an Artificial Intelligence–based natural gas leakage recognition system using thermal images. Thermal cameras capture infrared radiation differences caused by leaking gas, and deep learning models analyze these patterns for accurate detection. The system enhances safety by enabling realtime monitoring and automated alerts. Experimental results demonstrate improved accuracy and faster response compared to conventional methods. This approach ensures non-contact, continuous, and reliable gas leakage detection in realworld conditions