Lidar Point Cloud Data Point Mapping Method

With the laser scanner hardware and software technology continue to evolve, the accuracy of laser data are getting higher, which directly raises the quantity of points obtained for a certain area.So that, the volume of 3D point cloud data we obtained is growing, especially those obtained by laser scanner equipped on satellite, airborne and automotive, which can even reach more than 100 GB for the data file of this magnitude. Traditional drawing methods trend to decomposit a large file into several small files to be processed, which cause that the rendering speed slow, and not able to get the intuitive shape of the overall point cloud.As well,LOD rendering method, which ranges the point cloud data to generate a number of different resolution files, according to the viewpoint distance to call a different LOD layer to draw so that the speed has improved, but the LOD file takes up storage space.Point-based rendering method, greatly accelerated the model drawing, and does not require a lot of the preprocessed file. According to the characteristics of the airborne point cloud and Qsplat, a new large-quantity cloud data rendering method is raised here. This method is a combination of sub-node of the quadtree and octree point clouds of airborne, combined with Qsplat the data characteristics and the characteristics of the point cloud data storage, and Qsplat display ideas.This new method has the following advantage:1,Because of the combination of quadtree and octree, speed of sub-node and reading is higher, and the node level can be reduced to minimum, and cost can be lower;2, the preprocessed file does not need a lot of storage space;3, without loss of point cloud accuracy;4, shown in the perspective conversion effect is smooth.5, using memory-mapped, less demanding of computer memory.