Design Of Railway Clearance Intrusion Detection System Based On Smart Camera

With the rapid development of railway construction in China,the speed of the train is increasing gradually,and the problem of traffic safety is becoming more and more prominent and important.Accidents caused by the foreign body penetration of railway happen frequently,which has brought a great threat to the safe operation of the railway.The existing foreign body detection methods based on machine vision mostly adopt the way of embedded acquisition with PC image processing,which has the problem of poor real-time and high cost.For the longer lines,it can't be layout intensive along the railway.Therefore,it is necessary to develop a set of smart cameras and algorithms with image acquisition,processing and transmission functions in order to achieve faster real-time detection and processing,and alarm at a lower cost.Therefore,this thesis presents a detection system based on smart camera for railway foreign body penetration,and designs an algorithm based on smart camera for the foreign body intrusion detection.The functions of collecting,processing,analyzing,automatic alarm and foreign body image storage are realized by arranging distributed detection nodes along the railway.At the same time,the monitoring center has the ability to configure the parameters of the smart cameras for each detection node.The monitoring terminal can send configuration parameter of image processing algorithm to each detection node,and can be updated online.So that the intrusion detection algorithm can quickly adjust to the environment,which improves the practicability of the system.This thesis first introduces the overall design of the railway intrusion detection system based on smart camera,including the overall network structure of the railway intrusion detection system,composition and function of each module and the design of overall algorithm.Then,the hardware circuit design,the design basis,the selection of the main chip and the selection criteria of the smart camera are discussed in detail.The image processing algorithm based on DSP is designed,and the algorithm is programmed by C language in DSP.Finally,the basic functions such as image acquisition,image storage,image transmission,image processing,feature extraction,foreign body alarm and Ethernet transmission are realized.In order to improve the processing speed of the system,the system has been studied from the development of the program and the allocation of memory,and the memory allocation system and algorithm program have been properly designed.Transmission mode,storage structure,algorithm and the program are all optimized.For example,the single connected domain labeling algorithm proposed by this system can record multiple feature parameters while the connected domain is being marked.Thus the calculating efficiency is largely increased as a result.For the purpose to test the effect of the detection of foreign bodies in a smart camera,experiments were carried out in the campus and the railway field.The experimental results show that the average frame rate is 12 frames per second,which meets the requirement of real-time detection.The accuracy of the detection and alarm of the foreign body reaches 95.83%,which meets the requirement of accuracy.