Steel rail fastener abnormity detection method capable of automatically labeling samples

Abstract

The invention relates to a steel rail fastener abnormity detection method capable of automatically labeling samples. According to the method, the fastener samples can be automatically collected in thefastener area positioning stage to construct the training data set, and the training samples do not need to be manually collected and labeled. A multi-classification recognition model integrating a knowledge base obtained based on off-line learning and an on-line learning classifier is provided, and the adaptability of fastener classification to new line data is improved. According to the method,a template library is dynamically updated by utilizing the thought of online learning, so that the fastener area positioning module can adapt to track images of different railway lines or different sections. In addition, a deep convolutional neural network model is designed, multiple convolutional layers are used for extracting image features, the expression capability of the image features is higher, and the image classification precision can be effectively improved. Aiming at the problem of unbalanced sample number of different types of fasteners, the invention provides a random sorting strategy to reduce the influence of unbalanced sample number on the network performance.

Description

Technical field [0001] The invention relates to the technical field of rail transit safety, in particular to a method for detecting abnormalities of rail fasteners with automatic sample labeling. Background technique [0002] Fasteners are parts used to connect rails and sleepers on the track. Its function is to fix the rails on the sleepers, maintain the gauge and prevent the rails from moving vertically and horizontally relative to the sleepers. According to the different track plate types, the installation distance of the fasteners is between 550mm-650mm, which converts the number of fasteners per kilometer to about 3400 pairs, which corresponds to the total mileage of my country's high-speed rail lines. The number of fasteners on the line is huge. of. The rails are impacted by the lateral force of the train operation every day, which will cause changes in the fastening assembly structure, resulting in failure of the buckling and displacement of the rail, and further major saf...

AI Technical Summary

Problems solved by technology

[0011] When the existing supervised railway fastener defect detection method detects new railway lines, it needs to re-collect and mark a large number of samples for training, which has the problem of poor promotion ability

Images