Cross rod end fastening bolt loss fault detection method based on deep learning

Abstract

A cross rod end fastening bolt loss fault detection method based on deep learning relates to the technical field of freight train detection, and aims to solve the problems of low efficiency in detecting the failure of the failure of the fastening bolt at the end of the cross bar, high labor cost and high time cost in the prior art. The beneficial effects of the invention are that: 1, a deep learning model is adopted to detect a cross rod end fastening bolt loss fault, fault detection efficiency can be improved, manpower cost and time cost are reduced, especially railway freight car images areall shot in the complex environment, a traditional image algorithm has high image requirements and certain limitations, and the deep learning detection mode can better adapt to fault detection of thecomplex images; and 2, a TensorRT acceleration optimization deep learning model structure is adopted, and a parallel detection image algorithm is designed, so that the fault detection speed is greatlyimproved, and real-time fault detection and alarm of real-time train passing of a railway can be met;

Description

technical field [0001] The invention relates to the technical field of freight train detection, in particular to a deep learning-based method for detecting the loss of fastening bolts at the end of a cross bar. Background technique [0002] As an important part of the whole railway 5T system, the Freight Car Fault Trackside Image Detection System (TFDS) plays an important role in preventing the faults of freight cars. , Misjudgment phenomenon also occurs from time to time, so it is urgent to grasp the characteristics of frequent faults and typical faults to improve the efficiency of vehicle inspection faults. [0003] The fastening bolt at the end of the cross bar is an important locking device for fixing the cross bar of the bogie at the bottom. During the operation of the truck, the bolts at the end of the cross bar will gradually loosen with the vibration of the cross bar, causing the position of the locking plate to change. If the deflection is serious, the locking plat...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to propose a deep learning-based fault detection for the loss of fastening bolts at the end of crossbars in view of the problems of low efficiency in detection of missing fastening bolts at the end of crossbars and high labor and time costs in the prior art method

Images