Abstract: Aiming at technical problems of long index construction time and large memory usage in existing massive point cloud visualization method, a fast visualization method based on octree and OSG paging nodes is proposed, which is rapid on the establishment of point cloud index and real-time in scheduling based on small memory usage. Using octree data index structure for data organization of point cloud, this paper establishes the level of the octree node file and saves block in file mapping way. The file node is reorganized into multi-resolution point cloud to support the OSG rendering engine. Using real-time scheduling technology based on OSG paging nodes, mass cloud points are highly quality visualized. Compared with the two mainstream point cloud processing commercial software, the results show that this method has the advantages of fast index establishment and small memory occupation, and interactive visualization is more smooth, which is suitable for mass data scheduling management and real-time visualization of various computer configuration.

Key words: octree, uniform sampling, OSG paging node, massive point cloud, fast visualization

摘要： 针对现有海量点云可视化方法存在索引构建时间长、内存占用大等问题，研究一种八叉树索引结合OSG分页结点的快速可视化方法，可在占用较小内存的基础上快速建立点云索引并实时调度。采用八叉树索引结构对海量点云进行数据组织，建立各层级的八叉树结点并以文件映射的方式分块保存，对结点文件重组织转换为支持OSG渲染引擎的多分辨率点云数据。采用基于OSG分页结点的实时调度技术，对海量点云进行高质量可视化。与目前两款主流的点云数据处理商业软件进行实验对比分析，结果表明所提方法具有索引建立速度快、内存占用小等优点，同时可视化交互更加流畅，适用于各种配置计算机下海量点云数据的调度管理与实时可视化。

关键词: 八叉树, 均匀采样, OSG分页结点, 海量点云, 快速可视化