| Safety is the eternal subject of railway transportation.Rail is an important part of tracks,broken rail may cause train derailment,causing serious safety accidents.Monitoring the broken rail can effectively improve the safety of railway transportation.A scheme for monitoring broken rail by detecting rail cracks was proposed in this thesis.The morphological characteristics of rail cracks were analyzed,and the existing crack detection algorithm was improved.The broken rail remote monitoring system based on image recognition was designed,implemented and tested.The main work of this thesis includes the following aspects:Firstly,the rail crack image recognition algorithm based on image processing was studied,and the crack detection algorithm based on morphological operation and the crack detection algorithm based on texture fusion and image saliency detection were selected for comparison.And two algorithms were improved according to the characteristics of the actual rail fracture scene,Secondly,the overall design scheme of broken rail remote monitoring system based on image recognition was carried out.Based on the analysis of system functional requirements and performance requirements,the system structure was divided,the hardware was selected,the system software structure was designed,the data flow in the software was analyzed,and finally the database was designed.Thirdly,the broken rail remote monitoring system based on image recognition was implemented,mainly includes:1)The system hardware and network environment were built,and rail monitoring video data collection and transmission function were completed;2)The Web client software was implemented,and functions such as data display,video streaming,and ROI drawing were completed;3)The server software was implemented,the rail crack detection algorithm was implemented by using Open CV,and database management,video stream transcoding and other functions were also completed.Finally,the performance test of rail crack detection algorithm and the function test of the system were carried out.100 rail pictures were selected as test samples to test the two rail crack detection algorithms,and the test results show that: the improved rail crack detection algorithm based on morphological operation has better results,so the improved rail detection algorithm based on morphological operation was selected as the crack detection algorithm of this system.The function of the system is tested,the test results reached expectations,indicating that the scheme can meet the requirements of the broken rail detection system. |
PDF Full Text Request