Research On The Rail Wear Measurement System Based On Machine Vision

Rail wear is one of the main defects of the track infrastructure.Accurate measurement of rail wear is of great significance for rail maintenance and train operation safety,With the rapid development of the rail transportation industry,dynamic measurement can’t be realized with traditional manual contact measurement,which is inefficient and cannot meet the demands of the maintenance department.Therefore,a non-contact rail wear measurement system installed on the track inspection trolley is developed based on line structured-light vision measurement technology,to achieve dynamic and accurate rail wear measurement.First of all,the current research of rail wear measurement technology is analyzed in this thesis,and non-contact method based on line structured-light vision is determined.The overall scheme of the system and crucial hardware are designed on this basis,and a measurement platform is established.By studying the imaging model and calibration methods of camera,the mathematical model of the line structured-light measurement system is established,and the calibration schemes of camera and structured-light plane are determined.The system calibration accuracy is verified by experiments.Next,the processing algorithm for rail structured-light image is designed,including filtering,threshold segmentation,center extraction and coordinates recovery.The improved Otsu algorithm is proposed for the threshold segmentation of non-uniform structured-light image.Moreover,the computational efficiency of the classical Steger center extraction algorithm is improved.The validity and efficiency of the improved algorithms are verified by experiments.Thirdly,registration of rail profile which is important in dynamic measurement is studied,and a two-stage registration method based on RDP and improved ICP algorithm is proposed.At the end,with the help of OpenCV and Qt open source libraries,the rail wear measurement system software and interactive interface are developed on Visual Studio 2015 platform,which can realize image acquisition,wear measurement and data management.In order to verify the system accuracy and efficiency,static and dynamic experiments are carried out in the laboratory environment and actual railway.The results show that the system measurement accuracy can reach ±0.2mm,and the efficiency is high,which can meet the requirement of rail wear measurement installed on the track inspection trolley.