Research On Key Technology Of Intelligent Identification Of Train Comprehensive Circuit Information

The structure of the train electrical system is different from the mechanical structure,and its complicated structure is prone to failure.The fault diagnosis and maintenance of the train electrical system affect the normal operation of the entire rail transit network.At present,the fault diagnosis and overhaul of the train electrical system depends on the paper version of the train's comprehensive circuit diagram,so there are problems such as inconvenience to find the original data and difficulty in overhaul,etc.,which reduces the work efficiency.Based on this situation,domestic subway companies have developed simulation software for the integrated circuit of the train electrical system,but this software requires manual collection and identification of the element information of the integrated train map and entry into the software system,which results in The workload of extracting the elements of the train comprehensive route map will be very large,and it will be limited by factors such as equipment,working environment and human subjective influence.Therefore,the identification of the comprehensive train map is very important.How to quickly and non-humanly identify it to improve the identification efficiency and shorten the identification time is a problem that needs to be solved urgently.In view of the above problems,the main research contents of this article are as follows:(1)Analyze the current technology of manual collection of integrated circuit information and the method of identifying integrated circuit diagrams,and find out the shortcomings and limitations of this manual identification method,and combine the element information of the integrated train circuit diagram with the existing The theory of image recognition is suitable for the intelligent recognition method in this paper.(2)Analyze the problems that the paper version of the train's comprehensive route map in the actual project is easily damaged and unclear to upload pictures,and the traditional single image recognition method is not accurate.The processing technology performs image preprocessing on the comprehensive train map.A hybrid recognition method is adopted for the preprocessed pictures.First,the template matching method is used to identify the characters,the components are windowed,and the line path is identified by the area tracking method.(3)Analyze the genetic algorithm and BP neural network algorithm,design a genetic network algorithm suitable for the intelligent recognition of the train comprehensive route map,and establish an intelligent recognition model.Use the results of template matching and recognition as model input layer information and set up a classifier to match and select the input information,and then train the classified results with the three sample sets of established characters,components and line paths,and intelligent recognition process Binary coding is used to represent the identified elements,and the results of several recognition methods are compared through experiments.(4)Design and develop software for intelligent identification system of train integrated circuit information,use Java Script,HTML5 + CSS3 and other languages for programming,and use SQL Sever database and other software as system development tools The environment,framework,modules and functions of the developed system are introduced.The developed system is tested in terms of performance and compatibility,and the developed system is applied to the actual fault diagnosis and maintenance of the train electrical system and In the actual project of the subway company's study and training.