Research On Organization Andvisualization Of Massive 3D Laserpoint Cloud Data

In the mid 1990 s, the ground laser scanning technology began to be widely used in the modeling of complex scenes and spatial entities. The massive point cloud data obtained through this technology is an important information source in the threedimensional space coordinate system, which plays an important role in the marine surveying, geographic information system and digital city construction. Therefore, how to use the existing computer processing power, the massive point cloud data were effectively organize and index, more quickly and accurately complete point cloud data of three-dimensional visualization modeling has become an important research topic. At present, there are a lot of related research on point cloud data organization and management at home and abroad, and put forward the corresponding organization plan. The most common organization scheme of the combination of regular grid, traditional quadtree structure, R tree, KD tree and octree structure, different organizations have their respective advantages and disadvantages. Therefore, the most important is to find a suitable point cloud data characteristics of the organizational structure, to better improve the efficiency of the organization.The main work of this dissertation:(1) Analyses the domestic and foreign for lack of point cloud data organization scheme, for vehicle borne laser scanning system to obtain the point cloud data with the mass of random characteristics, puts forward the improved quadtree structure.(2) Put forward "the Hilbert- improved quadtree structure point cloud data. The structure based on the improved quadtree nodes order changing, makes sequence traversal of the quadtree nodes order completely accords with the characteristics of the Hilbert curve. To organize in order of the point cloud data, can effectively reduce the computer on while reading massive point cloud data I/O interactions, Di Gao Dian cloud data spatial index efficiency; At the same time using the Hilbert curve reorganize the quadtree, will the data into multiple single resolution data.(3) Using DME technology, and the thought of "divide and rule", this paper puts forward the "pyramid- PC" this kind of block hierarchical data model, using the Hilbert- improved quadtree structure processing every mass point cloud data, implement mass on-board laser point cloud data based on the viewpoint of LOD efficient visualization.(4) Finally, this paper design the verification experiment, the experimental results show that the improved organization indexing scheme has the rationality and validity. At the same time using Geo Magic Studio related visualization experiments, to verify the effectiveness of the LOD visualization solutions.At the end of the paper, the paper discussed how to make use of the existing computer resources to deal with the massive point cloud data more efficiently.