| With the increasing demand of railway transportation and the requirements of speed and stability in the construction of high-speed railway in our country,the requirements for the efficiency and accuracy of rail inspection are higher and higher.At present,most of the railway sites use manual measuring tools to detect the rail,and even rely on visual inspection to check the rail damage.The work efficiency of manual measurement tools and human eye observation is low,the labor intensity is high,the dependence on testing personnel is high,and the detection data is not convenient for digital management.The commonly used manual measuring tools use contact measurement method,which is easy to cause wear and tear at the contact point between the tool and the rail,so it is necessary to maintain and calibrate the tools from time to time.With the rapid development of computer technology,computer vision,photogrammetry and threedimensional structured light technology have been widely used in various industries.The use of non-contact close-range photogrammetry technology and three-dimensional structured light technology in rail detection can improve the efficiency and automation of rail detection and solve the problems of low efficiency,easy wear and inconvenient digital management of manual measuring tools.Therefore,the research of three-dimensional structured light rail detection based on close-range photogrammetry has important practical significance.This paper makes an in-depth study of close-range photogrammetry and three-dimensional structured light measurement,and analyzes the coordinate transformation in close-range photogrammetry and the relationship between image and object.Based on the mathematical model of collinear equation and coplanar equation,the process of using spatial rear rendezvous,image pair relative positioning and beam adjustment to calculate the three-dimensional coordinate information of coded and non-coded landmarks in rail stations is established.In this paper,the working principle of 3D structured light scanning system is analyzed,and a method to quickly and accurately obtain 3D point cloud data of continuous rail is proposed by combining close range photogrammetry with 3D structured light technology.According to the characteristics of the rail point cloud in the field,an impurity removal algorithm is designed,which can effectively remove the miscellaneous points near the rail point cloud,including large ground point clouds.A point cloud registration algorithm for rail is proposed.The principal component direction of rail waist plane is calculated by PCA algorithm,and the rough registration of rail point cloud and model point cloud is carried out.On the basis of rough registration,the target point cloud of precise registration is determined by k-d tree radius search,and then the fine matching of point cloud is realized by ICP algorithm.The railway field test shows that the three-dimensional structured light rail detection method based on close-range photogrammetry can accurately obtain the rail three-dimensional profile information,and the detection parameters can be calculated automatically from the rail three-dimensional point cloud.The accurate measurement of rail surface profile parameters can be realized. |
