Study on the Intelligent Fault Image Identification Algorithm for TFDS Freight Cars Based on Deep Learning
The accurate fault identification for key components of freight cars is very important for the operation of vehicles.The current identification mainly relies on the artificial analysis by dynamic car inspectors based on the images obtained by TFDS system,costing a lot of manpower and material resources,with lower analysis efficiency and great possibility of fault omissions,causing hidden risks to the running safety of vehicles.This article proposes a two-stage vehicle fault identification algorithm based on deep learning including the target detection and the component fault identification.Through model training of the marked images,it can realize the intelligent identification of vehicle faults.The effectiveness of this algorithm has been verified in Zhengzhou North Car Depot of China Railway Zhengzhou Group and it turns out that this algorithm can realize the accurate location,the automatic fault identification and the comprehensive result analysis of the components of vehicles,effectively improve the accuracy of fault identification of TFDS system,greatly reduce the omission rate,so it has an important significance for improving the intelligent level of the operation of freight cars and guaranteeing the operation safety.
NETL
NSTL
万方数据
维普
