Research Of Catenary Insulator Damage Detection Technology Based On Machine Vision

With the development of electrified railway in China,the increase of electrified railway construction and speed demand are increasing,so the requirements for the safe operation of electrified railway are becoming higher and higher.Insulator is an important component in the supporting and hanging device of electrified railway catenary,and its state is the premise of safe operation of electrified railway.Because the operating environment of the electrified railway catenary insulators needs to be exposed to the outdoors for a long time,and suffers from mechanical tension,electrical flashover,and long-term aging of the materials caused by the operation of the locomotive,insulators will be damaged to varying degrees.If the insulator is not detected and replaced in time,it will cause unforeseen accidents and huge economic losses.At present,traditional detection technologies require human-made real-time monitoring,and are inefficient and have poor accuracy.Therefore,this thesis introduces machine vision technology to the detection of catenary insulators.It is helpful to improve the detection efficiency and ensure the safe operation of electric locomotives by using machine vision technology instead of human eye to realize the intelligent detection of insulators.With the contact network insulator as the research object,combined with the machine vision related technology,in order to realize the automatic and efficient detection of the insulator of the electrified railway contact network,the insulator damage detection technology is deeply studied.If we want to realize the accurate detection of insulator status,we need to identify the insulator to be detected accurately.In recognition process,a binocular vision based recognition and location method is proposed.First,the feature matching method of template is used to detect and describe the feature of insulators by SURF algorithm,which is used to identify the target.On the basis of the target recognition,the parameters are obtained by the binocular vision principle and the camera calibration,and the 3D information of the target is restored.In the process of damage detection,in order to highlight the damage characteristics and improve the accuracy of detection,the wavelet transform principle is used to enhance the target,that is to use the principle of wavelet decomposition and reconstruction to decompose the target image multiscale,decompose the image into high and low different components,retain the low frequency image and carry on the Wiener Filtering to the high frequency sub-image.Wiener Filtering and wavelet threshold filtering are used to reduce the interference information,and then restructure each component to get the image clear edge information,and detect the damage according to the damage feature.In view of the catenary insulator damage detection problem such as recognition,positioning and detection,the method of recognition and location based on binocular stereo vision combined with SURF algorithm is proposed,and a method of wavelet transform enhancement is proposed,which combines the wavelet decomposition and the reconstruction characteristics to combine the Wiener filtering enhancement feature.Finally,combined with all the research contents,the visual recognition and detection system of catenary insulators is designed and implemented through Visual Studio 2013 and Open CV software platform.The system finally realizes the identification and positioning of the insulators,and completes the detection of the insulator strings and the cracks after the identification.Through the experimental test of the system,the correctness of the research content and the feasibility of the system are verified,which provides a theoretical basis for the study of the damage detection technology of the electrified railway contact insulators.