Research On Airbone LiDAR Point Cloud Data Filtering Algorithm Based On Terrain Simulation

Light Detection And Ranging is an active aerial remote sensing ground observation technology developed rapidly in recent years,which can obtain large-scale,high-density,high-precision 3D surface information.Although the hardware equipment of Light Detection And Ranging(LIDAR)is quite mature,the theoretical method for subsequent LIDAR point cloud data processing is still in the research stage,and many problems have not yet been solved.Point cloud filtering,which is the extraction of ground points,is a key step in airborne LIDAR data processing' and the premise and basis for subsequent feature classification.Therefore,point cloud filtering is one of the key issues in airborne LIDAR data processing.Traditional filtering algorithms are based on prior conditions and experience to set filtering parameters,which are fuzzy and uncertain.The optimal filtering parameters for different terrain features are different.The filtering effect of the same filtering parameter on different terrains is significantly different and it is difficult to meet the adaptive requirements.In this paper,the structure characteristics of airborne LIDAR point cloud data are analyzed,the classic filtering algorithm is studied in depth,and the main problems in the filtering process of the classic filtering algorithm are analyzed.In view of the above problems,the main research work of this paper is as follows:(1)Briefly introduced the structure and principle of airborne LIDAR system,summarized the difficult problems and research status of airborne LIDAR point cloud data filtering algorithm,and deeply studied the airbone LiDAR point cloud filtering algorithm based on terrain simulation.(2)Aiming at the problem that the grid parameter threshold of the classical Progressive TIN Densification filtering algorithm needs to be manually set according to the empirical threshold,a PTD filtering grid parameter determination method based on isolines is proposed.Using the isolines generated by irregular triangular net,the maximum building area is determined by the continuity and closure of isolines and the obvious height difference of the same edge area of the building to determine the specific parameter threshold of the grid,Then organize the point cloud according to this parameter.Relevant experiments show that the optimized PTD filtering algorithm has better filtering effect for complex terrain with buildings.(3)Aiming at the seed point gross error problem of classical moving surface filtering algorithm,a moving surface filtering algorithm based on multi-level seed point optimization is proposed.Firstly,grid index is established to divide the point cloud data into two-level grids,determine the first-level seed points,and use the first-level seed points to screen the second-level candidate seed points.When the number of seed points cannot meet the requirements,the surface growth of the seed points is carried out by taking the primary seed points as reference points.Finally,the selected seed points are used for surface fitting to calculate the true elevation of radar point cloud data and the elevation difference of the fitting elevation,and the adaptive elevation difference threshold considering terrain fluctuation is used to judge the ground points and non-ground points.Compared with the classical filtering algorithm,the results show that the filtering algorithm can effectively reduce three types of errors,has higher accuracy,and improve adaptive ability.