Visualization and Evaluation of Methane Gas Leakage by Thermal Image Processing Using Supervised Deep Learning Models

Abstract

An infrared (IR)/Thermal Camera is used to visualize gases that can not be detected by a visual camera or the naked eye. There are many benefits of using the thermal Imaging system for gas visualizing, such as a sooner spot of methane leakage and easier perceptive of how the plume moves and disperse. Imaging systems are capable of measuring the temperature of targets, especially solid bodies. This facility, however, is not good for gas temperature measurements. This is due to the complex study of infrared imaging physics and few difficulties associated with the procedures of radiation heat transfer. The goal of this work involves the usage of Deep Learning’s IR or thermal camera and image processing techniques to identify the presence and absence of methane gas leakage and to provide a consistent distribution of the present methane gas.