Development of image processing based contactless monitoring methods in pantograf-catenary systems

Nowadays, electric trains, an important mode of transport, have fast, reliable and comfortable structure. In addition to these advantages of rail transport, it is costly in terms of monitoring, maintenance, downtime and accidents. Therefore, monitoring detection algorithms with high accuracy for rail systems are developed.For a railway vehicle, the rails that wheels moves on and pantograph-catenary system that enables the transmission of electricity are the most important components. In this study, image processing-based monitoring and diagnostics methods for the pantograph-catenary systems have been developed. Especially, many failures can be occured in high-speed railways such as arcs, excessive friction, incorrect contact force, the top bar of the pantograph and catenary line failures. In the literature, however, there is contact and contactless tecniques for the diagnosis of faults, nowadays usage of the vision based contactless technical techniques has become mandatory. In this study, three main studies are proposed for vision-based monitoring and fault diagnosis on pantograph-catenary systems. The first study aims to determine the pantograph models by using image processing. For an accurate monitoring and fault detection it is important that determine the types of pantograph which can be used in various ways on trains. Here, the pantograph lines are determined by edge detection algorithms and the pantograph type is determined from the information of the position and angle of these lines. The second study aims to detect and monitor the point of contact of the pantograph. The catenery line has monitored the it navigates on the surface of the pantograph regularly by detecting the point of contant between the top bar of pantograph and catenery line detemined by image processing with proposed algorithms. In recent study, arcs at the point of contact can be detected too with images from thermal camera. Here, the inferences about the life of pantograph and failures can be ocur are obtained by monitoring upper region of pantograph continuously and also, proposed algorithms that aims to detect various failures are implemented on FPGAs. Consequently, in this study, image processing-based methods that aims to monitor and fault diagnosis of pantograph-catenary systems in the rail system with high accuracy have been developed and has been verified with real data. This study hes been supported by a research project in TUBITAK 1001 programme under Project No: 112E067 and the studies were concluded with various national and international publications.