Analýza a klasifikácia tramvajových vozov z optických senzorov

Abstract

The thesis deals with the identification of tram models from the utility signal. The thesis theoretically describes the popular monitoring methods in rail transport using industrial sensors and the various issues that need to be monitored. In more detail, the thesis describes the problems of flat wheels. The proposed algorithm for pattern identification serves as a module to extend the existing system for flat wheel detection, which was developed in cooperation between VŠB-TUO and DPO vehicle fleet. The algorithm is tested on data output from the aforementioned system. The thesis analyses the test results and discusses the limitations of the algorithm revealed by the tests.