Massive Three-dimensional Point Cloud Management And Visualization Research

With the construction of digital cities and the rapid development of large-scale three-dimensional data acquisition technology, three-dimensional laser scanning, aeros pace/aviation image intensive matching massive point cloud data, Most direct mani festation is the point of increasing density and more and more the number of point s. Using an Airborne/Car/terrestrial laser scanning system for data can up to te ns or even hundreds of G. The existing three-dimensional point cloud processing so ftware,such as SCENE5.0that is FARO’s Focus3D point cloud processing software, Cyclone6that is Leica’s HDS3D laser scanner supporting, Polywork,Geomagic and so on. Having their own strengths,But they still focus on dealing with the problem of modeling and poor support the mass point cloud. The reason is that the point cloud data organization and scheduling non-optimizedTo deal with these problem, the thiese focus on researching the the organizati on and spatial index of the massive three-dimensional point cloud data and analysis the current commonly used three-dimensional point cloud data space indexing meth ods on the basis of reasearching the obatain of the geographic scene3D point clou d data. Then the thises Proposed to improve the octree3D point cloud data organiz ation and,The paper presents improved octree3D point cloud data organization and index and use hybrid index to reduce memory consumption and improve the effici ency of query. On this basis research, Comprehensive utilization of memory file ma pping, visibility discrimination and multi-level LOD technology can reduce the num ber of point cloud drawn.The point cloud can also draw on an ordinary PC to achi eve fast and efficient point cloud darwn. On the basis of the theories and methods of research, massive3D point cloud visualization Prototype system was developed a nd verify the proposed algorithm effective. The main results can be summarized as follows:(1) The thesis research geographical scene3D point cloud data acquisition. Mainly t he obtain method have Airborne/Car/terrestrial3D laser scanning and aerospace/ground three-dimensional photographic image matching technique, The thesis eval uate the different access methods.The thesis summarize the acquired point cloud fil e unified into a binary form of this article needs. (2) The thesis research massive point cloud data organization and spatial index and analysis commonly used three-dimensional point cloud data indexing method and s ummarize it. An improved octree encoding scheme was proposed, On this basis, f urther proposed hybrid index KD tree leaf node data to reduce memory consumptio n and improve the retrieval efficiency.(3) On the basis of the massive three-dimensional point cloud data organization an d spatial index, when the point cloud was visualized, The thesis integrated use of memory file mapping, the judgment of visibility and the multi-level LOD technolog y to reduce the number of rendering point cloud. Our system can visualize the mas sive point cloud when use these optimal scheduling approach.(4) To verify the proposed algorithm, a massive three-dimensional point cloud visua lization prototype system have been developed to verify the effectiveness of the me thod.