Research And Development Of Massive Terrain Point Cloud Visualization And Processing System

After the "Digital Earth" strategy proposed, governments have risen the national digitalprocess to the height of the national development strategies. As a critical foundation of thenational digital process, the DEM data, received an unprecedented attention. How to get fastand efficient access to the DEM has also become an issue that can not be ignored. The late1980s, as an emerging space-to-earth observation technology, the airborne laser radarmeasurement technology was considered to be one of the most promising data source that canget fast and efficient access to high-precision, high-resolution DEM, and is now representingthe a new direction of development of the earth observation. The direct data airborne laserradar measurement techniques obtain is a massive point cloud, which needs to be treatedbefore we can get DEM. At the same time, the visualization of massive terrain point cloud hasa large number of applications in the field of GIS, virtual reality, simulation and games. Thus,the visualization and processing of massive terrain point cloud is a very important issue.In this thesis, we studied various algorithms of massive point cloud visualization andprocessing, together we analyzed their advantages. On this basis, we designed our algorithmsof massive point cloud visualization and processing. The visualization algorithms iscomposed of the LOD algorithm based on block hierarchical kd-tree, the simplified terrainremoving technology and the3D engine optimization. The point cloud processing algorithmsis composed of the automatic filtering algorithm that integrated clustering segmentation withsurface fitting, and the manual editing algorithm based on block space indexing.Based on these two algorithm, we in this article also develop a massive terrain pointcloud visualization and processing software system.The test results show that our massive terrain point cloud visualization algorithm cangreatly reduce the need for data drawn by the graphics hardware, thus significantly improvethe display speed; our point cloud filtering algorithm can obtain a more satisfactory filteringresults, and manually editing features can quickly carry out the deletion of the point cloud.