Research On The Application Of Point Cloud Adaptive Reduction In 3D Detection Of Rail Profile

With the continuous improvement of China’s high-speed railway network and the steady improvement of train speed,railway transport safety is facing greater challenges.As an important part of railway safety transportation,track detection and maintenance is becoming more and more important.At present,the traditional manual detection is still the main way of track detection in China,which is time-consuming and laborious,and it is difficult to meet the requirements of multi-parameters,high precision,high efficiency and digital management of track detection.With the increasing maturity of computer science,optical detection and sensor technology,there are a variety of ways to obtain track profile information.Structured light threedimensional imaging technology is a non-contact contour point cloud acquisition technology developed in recent years.It can quickly,accurately and completely collect the rail surface point cloud.The single point cloud has a wide acquisition range and can detect a variety of rail parameters at the same time,which greatly improves the accuracy and efficiency of rail detection.In this paper,the principle of structured light three-dimensional detection technology is studied,and on this basis,the original rail point cloud on the railway site is obtained.In order to solve the problems of large number of original point cloud and many miscellaneous points,and improve the detection efficiency and accuracy,an adaptive reduction algorithm of rail point cloud based on point distance is proposed in this paper.Firstly,the geometric characteristics and miscellaneous points’ s distribution characteristics of the original point cloud are comprehensively analyzed,and the original point cloud is straightened by using density clustering algorithm,plane segmentation algorithm,principal component analysis,normal direction and point cloud centroid position.Then,the miscellaneous points are removed according to the spatial position distribution of the point cloud and the rail profile information.Finally,the rail point cloud is segmented,and different voxel sampling parameters are set based on the point distance to simplify the point cloud at different locations.In order to accurately analyze the detection accuracy of the rail profile three-dimensional optical detection system studied in this paper,a detection accuracy inspection and analysis module based on standard gauge block is designed in this paper.it better overcomes the disadvantage that the true value of rail detection parameters is unknown,and provides an intuitive and effective evaluation method for the detection accuracy of this detection system.In order to verify the applicability and stability of the research content in this paper,experiments are carried out in different railway sites.the results show that this method can effectively simplify the rail point cloud and lay the foundation for efficient and accurate point cloud processing.At the same time,the results of the detection accuracy analysis module and the comparative analysis between the manual measurement data and the measurement data of the detection system also show that the detection accuracy of the three-dimensional detection system of rail profile meets the detection requirements of rail damage,and has reliability and feasibility.