Track fault detection method and system based on infrared thermal imaging and computer vision

Abstract

The invention discloses a track fault detection method and system based on infrared thermal imaging and computer vision. The method comprises the following steps that: an unmanned aerial vehicle carries out tram rail image collection; a ground station receives image data of a high-definition camera and image preprocessing is performed; multi-threshold rail region segmentation is performed twice on a dark region in a trench a bright region outside the trench and the rail region is segmented based on distance features of the adjacent bright and dark regions to extract a rail image; graying is performed on an infrared thermal image and a high temperature region on the rail is extracted based on a relative temperature difference method; the preprocessed image is superposed with a rail detection window and masking is performed to obtain a region of interest, and edge closing determination and filling are carried out on the region of interest to obtain a communication region, and the communication region is screened to obtain a suspected rail foreign object; the suspected rail foreign object is inputted into a BP neural network for identification to obtain a foreign object classification result. According to the invention, rail foreign object identification and temperature detection are performed in real time to reduce the accident occurrence rate of rail transit and improve the operation safety of the tram.

Description

technical field [0001] The invention relates to the technical field of track detection, in particular to a track fault detection method and system based on infrared thermal imaging and computer vision. Background technique [0002] Modern trams have entered people's lives with the characteristics of green environmental protection, safety, comfort, flexibility and convenience. However, compared with subways, modern trams do not have completely independent right of way. When the tram runs at a faster speed and has a larger passenger capacity, foreign objects on the track will pose a great threat to the safety of the tram. At present, the tram track fault detection technology mainly relies on manual operation detection and maintenance - slow detection speed, long time-consuming, low safety, high labor cost, tram track fault detection has even affected the daily operation of trams, trams The traffic on the track section is blocked, making the city traffic jam. Existing obstacl...

AI Technical Summary

Problems solved by technology

The obstacle detection technology based on computer vision mainly relies on the camera installed on the vehicle to obtain the image information in front of the vehicle, and uses digital image processing technology to detect obstacles. Although it has a certain early warning function for the detection area, the detection area is narrow and only It can detect obstacles while the vehicle is running, and cannot handle daily inspections when the track is idle and fault inspections in emergency situations

Images