A train track fastener defect detection method

Abstract

The invention provides a train track fastener defect detection method. The method comprises the following steps of constructing a Faster R-CNN network model, wherein the Faster R-CNN network model comprises a region generation network RPN and a fast region convolutional neural network Fast R-CNN, and the network initialization mode adopts pre-trained parameter initialization of similar tasks, andthen the Faster R-CNN network model is trained by using training set data; shooting and acquiring a fastener image on the subway line through photographing equipment; and inputting the fastener imageinto the trained Faster R-CNN network model, extracting a fastener area in the fastener image by the Faster R-CNN network model by using convolution operation and pooling operation, and performing defect type detection on the fastener area by using a loss function. According to the method, through the combination of the region generation network and the fast region convolutional neural network, the autonomous learning and the feature extraction of the training picture are carried out, so that fastener positioning and defect automatic detection can be carried out on the acquired large-size image, the intelligent degree is higher, the realization efficiency is higher, and the model applicability is stronger.

Description

technical field [0001] The invention relates to the technical field of train track defect detection, in particular to a detection method for train track fastener defects. Background technique [0002] Rail transit, as one of the most important modes of transportation, has achieved great development in the world, and the development in China is even more significant. The rapid development of railways has put forward higher requirements for transportation safety. The health status of rails and fasteners is the key to ensure safe and stable transportation of rails. Due to the contact friction between the wheel and rail, the impact of vibration and shock, and the influence of the natural environment, various diseases and defects are prone to occur on the track, such as damage and loss of fasteners. The function of fasteners is to fix the track on the track bed, but damaged or missing fasteners cannot play a safe and effective fixing role. The loss of the fasteners will lead to...

AI Technical Summary

Problems solved by technology

This processing method will cause the loss of image information, which may lead to poor subsequent recognition and classification results

In addition, a complete fastener is composed of different parts, and the partial feature extraction algorithm can only extract the global features of the image and ignore the local features

Images