Study On High-Speed OCS Dynamic Inspection System Based On Image Processing

The contact line over the pantograph slides fast when the locomotion operates at high-speed. It will tent to cause power accident if the location of the line deviates too large. In order to assure that the lines and the locomotion operate normally, it is necessary to ensure the security of catenary and detect catenary geometry parameters frequently. Traditional detection approach tests data by the contact of the line and sensors installed on the pantograph. However, applying the traditional detection method on high-speed catenary testing will affect the following behaviors of pantograph and reduce the similarity between detection pantograph and actual pantograph on account of the vibration of high-speed catenary is always higher than the regular speed catenary, which results in the error of catenary detection parameters. Therefore, it is an urgent need to adopt new methods, new ideas for promoting the catenary detection technology development.This paper designs high-speed catenary dynamic parameters detection system based on image processing by tracking the trend of catenary dynamic detection domestic and overseas and considering the features about high-speed catenary detection fully. This system can achieve non-contact real-time detection of catenary geometric parameters by using Onuris linear array cameras installed on the locomotion to photogrammetry survey and applying optical triangulation principle and image processing technology.According to the distance of sampling points on space about the current frame is smaller than its image data before and after that frame, the value of image gray of adjacent frames and some other traits. The steps of image processing methods the paper used are as follows: firstly, frame of the current image and its adjacent frames before and after it are implemented the differential operation to eliminate the affection of uneven background; secondly, the contrast ratio between the contact line objective imagination and background can be improved by balancing image processing and enriching image information; thirdly, gray image first division can be realized and segmentation threshold can be determined by using Otsu algorithms, and fragmented interference objectives in image can be removed by domain division method; Finally, continuous of the contact line object image can be realized by using closed operation of mathematical morphology on binary image. In the paper, mathematics model of catenary objects forecast and tracking is built to remove interference objectives again and realize the computing of catenary geometric parameters. According to the character of locomotive running and factor of detection systems itself, the paper analyses the reasons for the system error and puts forward the corresponding compensation methods.Hardware architecture and holistic designation of software of the detection system are discussed systematically in the paper. The rationality and feasibility of system's designation is proved by debugging operation.