Machine vision based automatic defect detection method for train tunnel cable clamp

Abstract

The invention discloses a machine vision based automatic defect detection method for a train tunnel cable clamp, and relates to the field of image processing, especially cable clamp image processing in a train tunnel. The method comprises: first, collecting a clamp image; performing a vertical projection on the image, and determining whether a pipe exists in the image according to a protection curve, if yes, determining the location of the pipe and intercepting a pipe image; performing histogram statistics, performing inverting binarization according to a low threshold, and then determining whether a clamp exists on the pipe of the pipe image, if yes, determining a location of the clamp; and performing histogram statistics on the intercepted image, performing inverting binarization according to a high threshold, and then determining whether the clamp on the pipe of the pipe image falls. According to the method, a clamp that falls can be found conveniently, quickly and efficiently, so that the clamp can be replaced timely.

Description

technical field [0001] The invention relates to the field of image processing, in particular to the image processing of cable clamps in train tunnels. Background technique [0002] The cable clamp is to prevent the cable from moving due to the external force suffix self-weight after laying, to eliminate the eddy current phenomenon, and to prevent the jumping and displacement of the cable due to the generation of electric force. Therefore, it is necessary to use a clamp to fix the cable in sections. Due to the complex installation environment of the cable clamps in the train tunnel, the cards on the cable clamps will age and fall off for a long time, and the fallen clamps need to be replaced. At present, my country's train mileage is long, and the tunnel sections that need to be inspected are many and complicated. At present, my country still uses the method of manual visual inspection to detect the falling off fixtures. Not only the detection accuracy is low, the speed is sl...

AI Technical Summary

Problems solved by technology

At present, my country's train mileage is long, and the tunnel sections that need to be inspected are many and complicated. At present, my country still uses the method of manual visual inspection to detect the falling off fixtures. Not only the detection accuracy is low, the speed is slow, and the cost is high, but also it is detected by humans on the railway It is not safe, how to detect the clamp that has fallen off the card on the clamp at the lowest cost, faster, and with higher accuracy has become an important issue in cable maintenance

Images