Lidar Based Railway Clearance Detection

Railway clearance detection is one of the most important tasks to ensure the safety of railway operation.Recently,Laser scanning is a new emerging clearance detection technique,and attracks many researchers' interest.This paper mainly focuses on the key issues of Lidar-based railway clearance model.Our works can improve the efficiency of railway inspection tasks.Our contributions are as follows.(1)We proposed a model compound method based on graphics cropping in computer graphics.Firstly,for the simple clearance model,we proposed a simple and effective data structure,which can easily realize the generation algorithm of the clearance model.The algorithm can preserve the lattice data and vector data of the clearance model.For the existing lattice points,we proposed a vectorization method,which transforms the clearance model into vector format and facilitates the realization of the subsequent intrusion algorithm.Secondly,on the basis of simple clearance model,we proposed an effective algorithm for cutting polygons,which can achieve the merging of multiple clearance model and greatly improves the efficiency of the clearance model.(2)We proposed a real-time rendering method based on CUD A and OpenGL interoperation.This method greatly shortens the drawing time of point cloud data and realizes the accelerated rendering of point cloud data.Before improvement,we use the QwtPlot3D library to render the data,with the increase in the amount of data,drawing speed can not meet the requirements of the drawing.In this paper,we analyzed the QwtPlot3D library source code and determined the main reason for the large program overhead.And then we propose a method based on CUDA and OpenGL interoperation.And for the different data,we have experimentally compared this two methods,experimental results show that,compared with QwtPlot3D library,the new method greatly shortens the time of data rendering.