Method for carrying out defect identification on overhead line system image of railway

Abstract

The invention provides a method for carrying out defect identification on an overhead line system image of a railway. The method comprises the steps: photographing a railway contact network, acquiringan image data set of the railway contact network, and performing expansion processing on the image data set; constructing an improved Faster R-CNN model based on image pyramid feature fusion, and training the improved Faster R-CNN model by using the expanded image data set to obtain a trained improved Faster R-CNN model; and performing defect target identification on the to-be-identified image data of the overhead line system by using the trained Faster R-CNN model through a target detection algorithm to obtain a defect target in the image data. According to the invention, the improved FasterR-CNN algorithm is applied to defect detection of the image data of the railway overhead line system, automatic identification of overhead line system part defects is achieved, and the detection efficiency is improved. The unmanned aerial vehicle is used for shooting the railway overhead line system, overhead line system remote sensing images with better image quality can be obtained under the condition that line operation is not affected, and incomplete data caused by mutual shielding of the overhead line systems can be avoided.

Description

technical field [0001] The invention relates to the technical field of image data processing, in particular to a method for identifying defects on catenary images of railways. Background technique [0002] The catenary is the key equipment to ensure the safe operation of the train. Due to repeated agitation and vibration during train operation, catenary components are easily damaged or even lost. At present, the detection method of catenary is mainly to manually read a large amount of image data offline. However, with the large-scale construction of high-speed electrified railways, the number of artificial visual inspection photos is huge and the detection efficiency is low. The different cameras mounted on the inspection vehicle usually shoot at night, resulting in poor image quality and omissions. By using drones to shoot railway catenary, obtaining high-quality catenary images can avoid missed inspections caused by shooting angles. Using the deep learning target detec...

AI Technical Summary

Problems solved by technology

[0003] In recent years, with the development of graphics processing technology, some researchers have proposed traditional image-based detection methods to detect defects in the image data of the railway catenary. Although the appearance defects of the railway catenary can be detected to a certain extent, the algorithm is limited by the surrounding The environment has a large impact, and the accuracy and speed have not met the ideal requirements

Images