OpenMP并行算法在卫星重力场模型反演中的应用  

APPLICATION OF PARALLEL ALGORITHMS BASED ON OpenMP TO SATELLITE GRAVITY FIELD RECOVERY

作　　者：周浩[1] 钟波[1,2] 罗志才[1,2] 张坤[1] 

机构地区：[1]武汉大学测绘学院,武汉430079 [2]地球空间环境与大地测量教育部重点实验室,武汉430079

出　　处：《大地测量与地球动力学》2011年第5期123-127,共5页Journal of Geodesy and Geodynamics

基　　金：国家自然科学基金(40874002;41104014);国家863计划项目(2008AA12Z105);测绘遥感信息工程国家重点实验室专项科研经费资助项目(2009);地球空间坏境与大地测量教育部重点实验室开放基金(10-02-09)

摘　　要：利用卫星重力数据反演地球重力场需要解决重力场模型的高效计算问题。分析了最小二乘直接法求解重力场模型涉及的密集型计算任务,基于OpenMP实现了卫星重力场模型直接求解的并行算法。利用30天、5秒采样间隔的沿轨扰动位T和径向扰动重力梯度Trr数据,分别反演了60阶次的卫星重力场模型,计算结果表明,OpenMP并行算法能够有效提高直接法求解卫星重力场模型的计算效率,并具有很好的稳定性。The massive satellite data presents a significant computational challenge to estimate the gravity field.Analyzing the computing-intensive tasks in recovering the earth gravity field model based on the direct least-squares algorithms,on applying parallel computing techniques capable of rigorously inverting geopotential coefficients using direct method on the basis of OpenMP is focused.Eventually,the earth’s gravity fields complete to degree and order 60 are recovered respectively on the basis of monthly along track disturbing potential observations and gravity gradient observations（with 5 s sampling interval）.The simulation shows that introducing the parallel algorithms based on OpenMP into determination of the earth’s gravity field model can be more effective,and is of great reliability.

关 键 词：并行算法 重力场模型 卫星重力 重力梯度 OPENMP 

分 类 号：P223[天文地球—大地测量学与测量工程] P228[天文地球—测绘科学与技术]