Pantograph carbon contact strip abrasion detection method based on deep learning target detection

Abstract

The invention discloses a pantograph carbon contact strip abrasion detection method based on deep learning target detection. The method comprises the following steps of making an original image data set, and manually calibrating the position of an original image carbon contact strip; clustering the marked bounding boxes in the data set by using an unsupervised learning algorithm k-means to obtain Anchor Box of which the size and shape meet requirements, and training a Yolo model under a deep learning daknetk framework to obtain a pantograph carbon contact strip positioning model; determining a rectangular region completely including the pantograph carbon contact strip by using the pantograph carbon contact strip positioning model, and intercepting the coordinates of the rectangular region in an original image; and extracting the image edges by using an adaptive threshold edge detection algorithm, determining the minimum distance between the upper and lower boundaries of the carbon contact strip by using a projection method, and calculating the thickness of the carbon contact strip. The method can adapt to a complex environment, the positioning rate is improved, and the robustness and the accuracy of a pantograph carbon contact strip abrasion detection algorithm are improved.

Description

Technical field [0001] The present invention relates to the technical field of traffic safety engineering, in particular, a substantially called carbon slide wear detection method for deep learning target detection. Background technique [0002] With the rapid development of rail transit in recent years, the driving safety of rail trains has received increasing importance. The strap is an electrical device that is mounted on the top of a power locomotive or electric vehicle group, from one or more contact wires. Direct contact with the contact wires, it can cause the electrical bow-based carbon slide wear to be thin. The thickness of the electric bow slide is too small, which not only affects the normal power supply of the urban rail train, but also the arc discharge thus generated further exacerbates the abrasion of the electrical bow slide and the contact line, so the abrasive detection of the electric bow carbon slide is the train. An indispensable part of overhaul, the thickn...

AI Technical Summary

Problems solved by technology

The pantograph wear detection method based on machine vision is to extract the edge of the pantograph carbon slide through the traditional image processing method. The traditional image processing algorithm has high requirements on the quality of the picture. The impact is large, and the positioning accuracy of the pantograph carbon slide is not high, so the accuracy of the wear detection of the pantograph carbon slide is low

[0004] Feng Yong from the Electrical Division of CRRC Qingdao Sifang Vehicle Research Institute Co., Ltd., Song Tianyuan and Qian Xueming from the School of Electronics and Information Engineering of Xi'an Jiaotong University proposed in the article "Rapid Detection Method for the Safety Status of High-speed Railway Power Receivers Based on Deep Learning" A fast and accurate method for detecting high-speed rail pantographs. This method first uses more than 10,000 pictures captured by the camera in front of the high-speed rail pantograph as training samples to generate a prediction model, and then uses the YOLOv2 algorithm to monitor the camera in real time. On-line detection of images can give early warnings of missing pantographs and serious deformations, but this method can only roughly locate the approximate position of the pantograph, and cannot accurately calculate the wear value of the carbon skateboard

Images