基于时间均衡的并行插值数据划分方法研究  

Data Partition Method for Parallel Interpolation Based on Time Balance

作　　者：钱辰[1] 窦万峰[1,2,3] 杨坤[1] 汤国安[3] 

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 [3]南京师范大学虚拟地理环境教育部重点实验室,江苏南京210023

出　　处：《地理与地理信息科学》2013年第4期86-90,F0002,共6页Geography and Geo-Information Science

基　　金：国家自然科学基金项目(41171298);国家863计划项目(2011AA120303)

摘　　要：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。该文提出了均匀分布下基于时间均衡的数据划分方法。针对面积均衡的初始划分进行处理过程的时间分析,讨论了在均匀分布的假设下搜索时间与点密度的关系,并给出全过程的时间估算方法;通过时间估算得出时间系数,依据时间系数进行再次划分,从而实现基于时间均衡的静态划分。通过对比串行和并行计算的执行时间,以及与传统的数据划分方法下并行插值的执行时间相比较,分别验证了该方法的实用性和高效性。It consumes a lot of time to generate an interpolated DEM when discreted point cloud data is magnitude. Interpolation calculation to which parallel computing is applied can significantly shorten the calculation time. The current research achievements on data partitioning in parallel interpolation are summarized. A data dividing method based on time balance which is easy to implement is proposed in this paper. The paper first discusses the relationship between the searching time and point density under the assumption of uniform distribution. Then a whole process time estimation method is given. Using the method time coefficient can be calculated, Finally,according to the time coefficient the initial partitions are re-divided again so as to realize the static partitioning based on time balance. By comparing the calculation time of serial with parallel computing and the executing time of this method with traditional method, the effectiveness efficiency of the method are verified by the experiments respectively.

关 键 词：数字高程模型 并行插值 点云 数据划分 时间均衡 

分 类 号：P208[天文地球—地图制图学与地理信息工程]