面向离散点云并行插值数据划分方法研究  

Research on Data Partition Method for Parallel Interpolation of Discrete Point Cloud

作　　者：钱辰[1] 窦万峰[1,2] 

机构地区：[1]南京师范大学计算机科学与技术学院,江苏南京210023 [2]江苏省信息安全保密技术工程研究中心,江苏南京210097

出　　处：《南京师范大学学报（工程技术版）》2013年第2期63-67,共5页Journal of Nanjing Normal University(Engineering and Technology Edition)

基　　金：国家自然科学基金(41171298);国家"863"重点项目(2011AA120304)

摘　　要：插值是数字高程模型的核心分析方法,也是构建数字高程模型的常用手段.面对海量的离散点云数据,插值生成DEM的过程需要消耗大量的时间,将并行计算运用到插值计算中会显著缩短计算时间.结合点云数据分布不均匀的特点,本文提出了一套针对离散点云并行插值生成格网DEM的数据划分方法,将读取、搜索邻域、插值计算、输出4个过程的处理时间量化,保证划分后形成的子块处理时间均衡,提高了并行计算效率.Interpolation is the core analysis method of the digital elevation model,which is also a common means of the con- struction of the digital elevation model. It consumes a lot of time to generate an interpolated DEM when discrete point cloud data is magnitude. Applied parallel computing to interpolation calculation can significantly shorten the calculation time. Taking the point cloud data distribution characteristics into consideration,this paper raises a data partition method, which is used for parallel interpolation of discrete cloud points to generate grid DEM. This approach quantifies the processing time of four procedures, such as input, searching neighborhood, interpolation calculation and output. This method can keep the balance of the processing time of the sub-blocks thereby improving the efficiency of parallel interpolation.

关 键 词：数字高程模型 并行插值 离散点云 数据划分 

分 类 号：TP301.5[自动化与计算机技术—计算机系统结构]