Theory And Method On Point Cloud Data Spatial Management Based On Kd-tree

With the development of computer technology and the constant increasing of market demands, laser scanning technology, as a method of obtaining 3-dimensions space information in real time, is developed very fast. The laser scanning system has broad the source of data effectively, convert the data acquisition method, and can get higher-resolution digital surface model more easily. However, the hugeness of point cloud data always is the important factor which restrict the development of data processing method. It is urgent needed to find a spatial index to manage point cloud. Therefore, it is meaningful to study spatial index method and data processing method, the theory of improving data quality and processing ability of point cloud.In this paper, point cloud getting theory, point cloud index method and point cloud filter method are concerned. The main content includes:1. Systematically reviews the development course, relevance study current situation and application current situation of LIDAR and processing point cloud.2. The working principle, system components, system characteristic of Airborne Laser Scanning and Terrestrial Laser Scanning , and also the difference between LIDAR and INSAR, LIDAR and photogrammetry, are systematically analyzed.3. Based on the characteristic of the point cloud structure, the paper works out the disadvantage where the point cloud structure is managed by regular grid or octree index method.4. By analyzing the kd-tree index structure and retrieval efficiency, we comes to that kd-tree is a suited index method of point cloud data for the management. For the distribution of mass and non-uniformity characteristics of point cloud data, we improve the ordinary kd-tree storage methods and delineation of the rules, research and design the improved structure algorithm of kd-tree structure.5. Several important filter arithmetic are compared and analyzed, and then a new improved segmenting slop-distance filter based on moving round window filter method on the basis of slope based filtering arithmetic is brought out. Some experiments are carried on using point cloud data, and DTM is derived successfully.6. System structure of point data processing software is designed and he point cloud processing software is developed based on the research results of above theories.