海量低空机载LiDAR点云的地形断面快速生成算法  

A Rapid Algorithm for Terrain Profile Generation Based on Low-Altitude Airborne LiDAR Point Clouds

作　　者：周建红 杨彪[2] 王华 张行南[2] 蒋建平 李浩[2] ZHOU Jianhong;YANG Biao;WANG Hua;ZHANG Xingnan;JIANG Jianping;LI Hao(Hydrology Bureau of Changjiang Water Resources Commission,Wuhan 430010,China;Hohai University,Nanjing 210098,China;Lower Changjiang River Bureau of Hydrological and Water Resources Survey,Nanjing 210011,China)

机构地区：[1]长江水利委员会水文局,湖北武汉430010 [2]河海大学,江苏南京210098 [3]长江下游水文水资源勘测局,江苏南京210011

出　　处：《测绘科学技术学报》2018年第2期170-174,共5页Journal of Geomatics Science and Technology

基　　金：国家自然科学基金项目(51420125014)

摘　　要：针对现有算法二次插值造成的精度损失问题,结合新兴的低空机载LiDAR技术,提出了一套高精度地形断面快速、自动生成算法。该算法通过内存映射和断面线缓冲区分析,从海量点云数据中快速提取构TIN点集;优化了逐点插入法,实现局部Delaunay三角网的快速构建;根据三角网拓扑关系,实现断面线与TIN的快速求交,生成并优化地形断面图。实验结果表明,算法克服了海量数据处理时计算机内存不足的瓶颈问题,并有效避免了二次插值造成的精度损失,且运行高效。In view of the accuracy loss problem caused by twice interpolation of the existing methods,a fast and automatic high-precision terrain profile generation algorithm based on newly-emerging low-altitude airborne LiDAR technique is proposed,which consists of the following steps. Firstly,TIN point set is extracted from the massive point cloud data quickly by memory mapping and section line buffer analysis. Then,the incremental insertion algorithm is optimized to realize the fast construction of the local Delaunay triangulation. Lastly,the terrain profile is generated and optimized by section line and TIN quick intersection according to topological relations between the triangles. The experimental results show that the algorithm not only overcomes the bottleneck problem of insufficient computer memory in massive data processing,but also effectively avoids the precision loss caused by twice interpolation and runs efficiently.

关 键 词：低空机载LiDAR 地形断面 点云 不规则三角网 DELAUNAY算法 

分 类 号：P237[天文地球—摄影测量与遥感]