面向海量精细点图像和散乱点云的多级混合空间索引机制

项目名称： 面向海量精细点图像和散乱点云的多级混合空间索引机制

项目编号： No.41301429

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 天文学、地球科学

项目作者： 郭明

作者单位： 北京建筑大学

项目金额： 25万元

中文摘要： 海量精细点云数据的空间索引设计不仅是点云组织和管理研究的主要内容，也是海量点云数据高效处理、分析与可视化的基础。本申请针对这一国际前沿的学术热点问题，以地面激光雷达获取的原始点图像和经预处理后大规模散乱点云为研究对象，提出了一套海量精细点云数据的多级分类空间索引方法，主要包括：基于八叉树空间索引和三维k-d 树的大规模散乱点云索引方法；针对两类海量点云数据的三维R 树多级混合空间索引方法；基于多分辨率细节层次和视窗裁剪技术的数据库数据检索和快速可视化算法。以.Net 和OpenGL 为开发工具，应用故宫博物院等实验区的海量多测站点图像和大规模散乱点云数据，设计开发实验原型系统进行算法验证。本研究成果有望形成一套海量精细点云数据组织的有效方法与关键算法，为点云数据的处理和分析提供技术支撑，同时推动地面激光雷达技术在文化遗产保护等领域的相关应用。

中文关键词： 三维空间索引；三维激光扫描；OctKDTree 树；曲面细分；三维空间可视化

英文摘要： The design of spatial index for huge fine point cloud is not only an important part of the organization and management work,but also the basis of efficient processing, analysis and visualization.Aiming to the international forefront academic issue for large-scale point-cloud data,an approach of multi-level hybrid spatial index is proposed in this proposal.For large-scale scattered point-cloud, a three dimensional integrated spatial index which are compose of octree spatial index and three-dimensional k-d tree index is used. Multi-station point image data and scattered point-clouds are organized by three-dimensional R-tree spatial index which are constituted by their MBBs. Combined with the view-window cutting and multi-resolution level of detail(LOD) technology, the database data retrieval and fast visualization of huge point-cloud will be implemented. The ground-based three-dimensional laser scanning data from Forbidden City are used to validate the feasibility and the effectiveness of management theory and methods above by the prototype system, using .net and OpenGL development tools. The results is expected to form a set of relevant theories, methods and key algorithms for large-scale point-cloud data organization.Effective support for management and application of point cloud data will be provided, while pro

英文关键词： three-dimensional spatial index；three-dimensional laser scanning；OctKDTree；Tessellation；three-dimensional spatial visualization