Complex Fuzzy Automata Based Fault Detection Approach Using Thermal Images in Railways

Abstract

One of the most important components for railway transport is the railway line and its components. Failures that can occur in the railway line and its components cause major accidents and financial losses. In this study, a new approach based on complex fuzzy automata is proposed to detect rail surface defects. The rail surface is detected using the thermal images taken from the rail line. Feature extraction is performed on the rail surface image and the obtained results are used as input values of the complex fuzzy automata system. With the inferences obtained at the output of Complex fuzzy automata, the condition monitoring is done on the rail surface. Recommended output of complex fuzzy automata; There are three different situations: solid rail, rail joint and defective rail. The proposed method shows that it is unique to study the difference between the rail joints and the defective rail surface. Moreover, the use of thermal images instead of normal images has increased the success of the method.