Research On Classification Method Of Point Cloud Data Based On Railway Re-measurement

The rapid development of China's railway Originated from Strong national strength,especially the massive construction of high-speed railway,the railway runs faster and faster and the increase of railway operation speed has put forward higher requirements on the safety and smoothness of train operation.Therefore,establishing a set of efficient,scientific and systematic railway operation and maintenance system is particularly important.However,intermittent re-measurement of the organization in railway operation and maintenance is an essential means to grasp the status quo of the railway and monitor the safety of railway operation.LiDAR can acquire large-scale and large-volume three-dimensional point cloud information.Compared with the traditional operation rehearsal,LiDAR has the advantages of fast,safe and high precision.However,there is huge amount of data and finding the right point cloud classification method is the key.The main work and research contents of this paper are as follows:(1)This paper introduces the practical problems to be solved urgently in the operation and maintenance of the rapid railway development in our country and discusses the advantages and challenges of the new LiDAR technology on existing railway lines.(2)A comprehensive introduction to the system components and working principle of LiDAR laser vehicle,access to literature,website for domestic and foreign research on this area of equipment.The characteristics of the data of laser radar measurement system and the progress of data processing technology are reviewed.The advantages and disadvantages of LiDAR laser technology are analyzed,and several main application fields of LiDAR laser technology are introduced.(3)Research on point cloud segmentation method: This paper discusses the deficiencies of several segmentation methods commonly used in vehicle laser LiDAR point cloud data processing.According to the requirement of railway re-measurement,the suitable segmentation method is adopted according to the distribution characteristics of the main features: for railroad tracks,the similarity measure of fusion reflection intensity is used tosegment the railroads according to the reflection intensity and characteristics of the railroad tracks;Point rate,the realization of the tunnel segmentation.(4)On the basis of point cloud segmentation,this paper summarizes the semantic features of the target point cloud and builds a knowledge base.The principal component analysis(PCA)method is used to extract the eigenvalues.On the basis of the eigenvalues,combined with the corresponding knowledge to achieve feature classification.