列车运行动态监测系统中故障图像识别算法的地研究.pdf

Abstract Fault Since Detection beenintouse 2000，Train System(TFDS)hasput in traindetectioninstitute camerais arrayplaced major gradually．High-speed of trackto train train atthesiderailway take of image keycomponents，SO car train faultfromthe taken．The identify inspector component image system makesthe oftrain shiftedfrOmfield toroom way inspection operation train modemeanslabor inspection intensity operation．However。theexisting isstill is infiuenced veryhigh，andgreatly bysubjective has to atrain faultauto-detection greatsignificancedesign component system whichcan currentfaultdetectionandlocationmethodandrealizethe improve shiftfromartificialdetectiontoautomatic way recognition． isdividedintofive TFDS faultauto-detection component system keysteps： extraction．featureselection、fault imagepreprocessing、feature recognition andfaultlocation．FirstOfall。some meanand preprocessingtechnique，the median andscale equalization filter,Laplaciansharpenin