Research On Light Band Thinning Of High-speed Turnout Wear Detection Based On Light-section Method

Wear of turnout is a common disease,and in order to ensure operational safety,the wear need to be measured accurately.At present,the monitoring system based on light-section method is the dominant.Processing the image of rail profile with advanced image processing technology will improve the accuracy of the rail wear monitoring effectively.Attention mainly be paid to the thinning operation of the rail profile image.The main works are as follows:Firstly,image acquisition subsystem of rail wear monitoring system based on light-section method was designed.After clarifying the theory of light-section method,the equipment that will be used in image acquisition subsystem were determined and chosen,mainly including CCD camera,lens and structural light source.In order to calculate the wear more easily and accurately,the spatial positions of the CCD camera and the rail,the structural light source and rail are designed.Secondly,the image preprocessing algorithms are analyzed,which mainly includs image filtering and image segmentation by reviewing the literature.In the part of image filtering algorithms,two basic algorithms,median filtering and mean filtering,are introduced.At the same time,a filtering algorithm based on image enhancement is discussed considering the actual situation that the image contrast and brightness in this study.In the image segmentation section,the edge detection segmentation and the threshold segmentation are discussed,and the Otsu algorithm is selected for the light band image segmentation at last.In addition,the light band image thinning method is studied.Based on the characteristics of image and thinning purpose,indexes of judging the quality of thinning are put forward.On the base of the analysis of the existing refinement algorithms,and considering the thinning quality and the operation cost of each thinning algorithm,Fast Parallel Thinning Algorithm is choosen as the fundamental algorithm for band image thinning.Aiming at the defects such as burr,bifurcation and non-single pixel width existed in the thinned image,the corresponding elimination templates are designed,optimizing the thinning algorithm to ensure the accuracy of the wear detection system based on the light-sectioning method.Finally,an image acquisition system is built in the laboratory environment,and the thinning quality of the image before and after algorithm optimization is verified.In order to carry on the thinning precision analysis,the automatic calculation of the rail wear is carried out,and the precision of the thinning is analyzed by the wear measurement result.It was proved that the optimized algorithm can effectively improve the precision of the light band thinning operation.Image thinning,which is based on light-section method,plays an important role in monitoring rail wear,and it is also one of the main research fields.Based on the study in this paper,the direction of future research was prospected.