Catenary Suspension Detection Identification Techollgy Research

In recent years, the rapid development of electrified high-speed electrified railway catenary equipment quality is particularly important, and gradually become the basis for the development of high-speed electrified railway. Catenary this detection technology research and application in the field has become the key to improving the quality of equipment and routine OCS defects overcome. To solve this problem, this paper outlines the catenary dynamic detection technique based on the historical development, compared to the non-contact detection system catenary and contact detection method advantages and disadvantages, image recognition system exposition of electrified railway construction and based detection system catenary detection system based on the composition, the system features, such as identifying ways and methods of computation analysis.This thesis includes five chapters, the first chapter introduces the background and significance of this study, and briefly describes the development and testing of domestic and foreign detect the status quo catenary catenary-based image processing. The second chapter introduces the principles of image processing and decision tree related algorithms. Grayscale image conversion, edge detection, Hough transform line segment extraction, sift feature matching and decision tree classification algorithm in detail the working principle and added some personal summary and experience. The third chapter introduces the paper design of identification and detection algorithms, and algorithm testing. Work process complete algorithm is first decomposed Catenary videos, all of its frames in sequence number marked on the map, then loop into the algorithm; then circulating in each image in turn gray transformation, edge extraction, Hough transform line segment extraction, classification rules are applied to the image generated by the decision tree to judge; finally judged to be stored in an array of image drawing number, according to an array of revised re-cycle call image, using sift feature templates fcr these images matching the image to identify the location of the locator, and then extracts an image locator in the tilt angle calculated by the geometric relationship between the slope of the positioner. The fourth chapter is summarized and analyzed the results of Chapter III of the test algorithm is designed mainly to show the third chapter, the algorithm speed and accuracy of data, and the algorithm is evaluated. The fifth chapter of the research results are summarized in this article and shortcomings, the next step for catenary Detection Based on Image Processing prospected.