摘要
利用八叉树数据结构对海量点云进行分块处理,将八叉树叶结点的点云逐层随机采样后保存在外存中构建多分辨率LOD数据结构,设计了一种基于视点的多分辨率点云内外存调度策略,实现了海量点云的流畅显示。通过对一组海量点云数据进行实验,分析了不同八叉树划分深度对八叉树划分、多分辨率数据构建以及显示的影响。
Smooth display of massive point cloud is the basis of point cloud data processing and analysis. This paper used octree data structure to block the massive point cloud. And then, in order to construct multi-resolution LOD data structure, the paper randomly sampled the octree leaf nodes of point cloud layer-by-layer, and kept the data in external storage. The paper designed a new method that in-core and out-of-core exchange strategy of multi-resolution point cloud based on the viewpoint, which could realize the smooth display of the massive point cloud. In the end, the experiments were carried out for a set of massive point cloud data, and analyzed the different octree partition of depth impacts on octree partition and generation of multi-resolution data and rendering. The result concludes that this approach in this paper is able to smoothly render hundreds of millions of points on the regular computer.
出处
《地理空间信息》
2016年第10期22-25,4,共4页
Geospatial Information
基金
国家自然科学基金资助项目(41371384
41401465)
地理信息工程国家重点实验室资助项目(SKLGIE2014-2-4-1)
关键词
八叉树
多分辨率LOD
海量点云
内外存调度
点云绘制
octree
multi-resolution LOD
massive point cloud
in-core and out-of-core exchange
point cloud rendering
