A Combined 2D and 3D Spatial Indexing of Very Large Point-cloud Data Sets

A database management algorithm based on combined 2D and 3D indexing of very large point-cloud data is proposed,for extracting the point cloud in need and improving the query efficiency.Single-station point-cloud is managed with 2D quad tree and 3D MBB structure.Multi-station point-clouds are indexed with 3D-R tree.Finally the organized hierarchical model and other attribute data are stored in ralation-object database.The data storage,management and visualization of very large point-clouds are implimented on personal computer with massive point clouds from the ancient buildings such as Forbidden City.Result shows that the algorithm is able to manage more than 10 GB-level data and one billion effective points with satisfactory drawing efficiency.