Dynamic Contact Position Of Pantograph And Catenary Based On Video Image Research On Real-time Monitoring Method

As a heavy power in high-speed railways,it is of great significance to the development of the entire society and economic progress.Ensuring the safety of high-speed railways is of great significance to the development of the entire train.Train pantographs are the key electrical equipment for trains to obtain electrical energy.Monitoring the train pantographs is critical to ensuring train safety.Maintaining good contact between the pantograph and the contacts of the catenary is the key to ensuring that the pantograph receives stable power from the catenary.The integrity of the horns ensures the stable operation of the pantograph.The carbon slide is the contact part of the catenary and pantograph.Its abrasion affects the stability of power supply.The current pantograph monitoring technology is difficult,and it is difficult to comprehensively monitor the pantograph.To solve this problem,this article conducts the following research:(1)Aiming at the problems of large amount of pantograph image data,complex background and feature extraction,the preprocessing method of pantograph video was studied.First,by separating the continuous video frame images from bow-net video image frames,key frames are extracted and key frame models are established.Then,in order to solve the complex problem of image background,K-means,superpixel,morphological watershed,and Grabcut algorithms were used to analyze and compare the effects of video processing on bownet video,so as to determine the background separation method suitable for the actual monitoring of bownet.On this basis,the LBP,Harris,and SURF algorithms are further used to extract the bow-net features,and the pantograph image features are extracted for each method and analyzed.(2)Aiming at the problem of dynamic monitoring of pantograph and catenary contact based on machine vision,four detection methods for bow net monitoring were studied.1.Real-time dynamic monitoring of the contact position of the pantograph and catenary.An improved Yolo-tiny-strong algorithm is proposed to effectively monitor the target of the contact position.2.The integrity of the pantograph horns is established based on the pantograph Mathematical model of the key point area characteristics of the shofar,which can quantitatively analyze and detect the deformation of the shorn;3.Arc burning phenomenon of the pantograph,study the positioning method of the arc burning zone of the pantograph and target identification monitoring method;4.Wear of the carbon slide of the pantograph,A detection method for quantitatively analyzing the wear of the pantograph's carbon slide by extracting the characteristics of the wear contour moment of the carbon slide is studied.(3)Set up a test platform for bownet video dynamic monitoring.Relying on the existing bow-net test bench in the laboratory,the CCD camera and Jetson TX2 hardware platform are used to complete the experimental platform construction and the environment configuration.Experiments were performed to verify the contact position,horn deformation,arc burning,and carbon slide wear monitoring and detection methods proposed in video image-based bow-net dynamic real-time monitoring.Through the acquisition,processing and analysis of video data,the experimental results obtained verify the feasibility and effectiveness of the bownet video dynamic monitoring method.Based on the video real-time monitoring method of pantograph and catenary contact position based on video images,the purpose of automatic monitoring of pantographs from the perspective of video images.It is very important for the track intelligent monitoring technology.In order to ensure the stable and safe operation of the train,an effective and feasible scheme is proposed,which can promote the smart track monitoring.