A Multi-Precision Partial KD Tree Index Based on Database

With the comprehensive application of airborne laser radar system, the data organization and management in three-dimensional point-cloud become a hot point in computer. This paper proposed a multi-precision partial KD tree index based on database to solve the problem on huge point-cloud data organization and management. This technique partition the point-cloud model into small disjoint blocks, in which the points are reorganized as a new KD tree, and build multi-precision partial KD tree something like Level of Detail Model. It´s introduced the concept of screen precision for point-cloud and its realization. The point data are written into database follow the layer of detail, Data of different level in database is transferred into memory by different view while rendering. Experimental results show that our approach can make a convenient management on huge point-cloud data and realize real-time rendering and roaming in large amount of point cloud data.