Designment And Application Of A Defect Identification Algorithm For Key Components In Straddle Monorail Transit System Based On Histogram Of Oriented Gradients Feature

With the expansion of urban scale,the increasing number of population and the increase of urban traffic congestion,it is more and more important to increase the urban rail transit lines and ensure the normal operation.As a mature urban rail transit tool,the straddle monorail has been popularized and built in many second-tier and third-tier cities in China because of the advantages of small noise,low cost and small effect on topography.Suspension support is the core component of the rigid catenary.Because of the special structure of the straddle monorail,there is a rigid catenary on both sides of the runway track,so in quantity,the support suspension device of the straddle monorail is two times more than the support suspension of the other rail transport tools,which is more important for the normal operation of the straddle monorail.The key components include support insulators,the bolts of insulators,the bolts of clamps and so on.The two kinds of bolts are not only the necessary components of suspension support,but also the important components for the stability of perpendicular offset.Therefore,it is necessary and practical significance to design a algorithm for identifying the two kinds of bolts of straddle monorail.This paper firstly proposes the concept of key frames.According to the size of the pictures,determining the range of key frames ensures the insulators locate in the middle of the pictures and the two types of bolts are located in the best position for observation and defect identification.By saving the key frame images,workers can only observe the key frame images with insulators and other components,filtering out many useless images,which improves work efficiency.The successful application of HOG feature and support vector machine in insulator location,abnormal state recognition and related fields is analysed.The method is used to locate the straddle monorail support insulators.If the position target rectangle is within the key frame determination range,the picture is considered as a key frame.In the key frame,the defect of two types of bolts are identified based on the method of coordinate determination and feature detection,and the system will record the type of defect.Finally,based on defect recognition algorithm of key components of straddle type monorail,relevant software was developed and successfully tested and applied in Chongqing straddle monorail actual lines.