双差定位程序反演算法剖析及改进  

Analysis and Improvement of Double-Difference Location Inversion Algorithm

作　　者：王清东[1] 王光明[1] 朱良保[1,2] 

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

出　　处：《大地测量与地球动力学》2016年第1期79-84,共6页Journal of Geodesy and Geodynamics

基　　金：国家自然科学基金(41374053);国家测绘地理信息局测绘基础研究基金(13-02-06)~~

摘　　要：详细分析双差定位程序中SVD和LSQR两种反演算法的优缺点,引入线性Tarantola反演法,通过动态数组和全局变量对程序进行改进,使程序结构体系更清楚,对计算机内存的要求降低,提高了计算效率。结果表明,双差定位法能够得到震源间精确的相对位置,Tarantola方法的定位精度和抗噪能力高于SVD方法,内存占用和计算速度也远远优于SVD方法。We analyze the advantages and disadvantages of the SVD and LSQR inversion algorithm in double-difference location program. The linear Tarantola inversion method, dynamic array and global variable are used to improve the program, making program structure clearer and easier to use, reduce requirements on the computer＇s memory, and improve the calculation efficiency. The theoretical ex- periment results and practical earthquake sequences location results show that double-difference loca- tion method can get accurate relative position between the hypocenter. The positioning accuracy and anti-noise capacity of Tarantola method is better than SVD method, and the memory requirement and computation speed of Tarantola method is also better than SVD method.

关 键 词：双差定位程序 反演算法 动态数组 全局变量 线性Tarantola反演法 

分 类 号：P315[天文地球—地震学]