| Interactive processing of 3D point cloud data generally depends on three-dimensional real-time rendering of point cloud data.With the further application of three-dimensional laser scanning technology,more and more three-dimensional point cloud data is acquired and accumulated.With the increasing amount of point cloud data,real-time rendering of large-scale three-dimensional laser scanning point cloud has become a bottleneck in the point cloud data processing.To deal with this problem,the paper presents a method that supports real-time rendering of massive point cloud.The main results are as follows.1.The point cloud data are all unified into the real geographic coordinate system by rotation and translation using the coordinate transformation matrix,contributing to the fusion of point cloud data from different scanning station through the registration.2.To be read directly Into the memory for processing,the large amounts of point cloud data is divided into a plurality of files,each containing a number of scanning points.3.The isolated points in each point cloud are filtered with the number of each point' s nearest neighbors as the threshold by constructing the kd tree and using PCL filtration.4.The spatial extent of the surveying area is computed by traversing the de-noised point cloud files and then the de-noised point cloud files are sequentially read and then octree index structure in different levels of detail is established on the external memory so that the mass cloud data can be organized effectively.The feasibility of the method is verified by comparing with the commercial software.Experimental results show that the proposed method supports real-time rendering of three-dimensional point cloud data whose volume is larger than the capacity of the physical memory.This method is expected for visualization and interactive processing of large-scale three-dimensional laser scanning point cloud data. |
PDF Full Text Request