基于选权迭代的总体最小二乘算法在三维坐标转换中的应用  

The Application of the Total Least Squares Algorithm Based on Reweighting Iteration to the Conversion of Three-Dimensional Coordinates

作　　者：徐波[1] 高井祥[1] 李增科[1] 刘洋[1] 

机构地区：[1]中国矿业大学国土环境与灾害监测国家测绘地理信息局重点实验室,徐州市大学路1号221116

出　　处：《大地测量与地球动力学》2015年第4期693-696,706,共5页Journal of Geodesy and Geodynamics

基　　金：国家自然科学基金(41074010);江苏省高校优势学科建设工程(SZBF2011-6-B35);新世纪优秀人才支持计划(NCET-13-1019);中国矿业大学基本科研业务费创新人才科研基金(2013RC16)

摘　　要：提出一种基于选权迭代的TLS算法,通过IGG I函数进行定权,降低含粗差观测值的权值,在进行三维坐标转换时能有效地处理观测值粗差。实例证明,该算法可以获取更高精度的三维坐标转换参数。Users of the least squares algorithm （LS） coordinate transformation, usually focus on ob- servation errors, but errors of the model coefficient matrix are ignored. Although total least squares algorithm （TLS） could take into account the errors of observations and the coefficient matrix, it cannot solve the problem of gross errors effectively if we start with coordinate data and observations with gross error. Through the IGG I function, which gives the right of the observation to reduce the weights of the observation with gross errors, this article presents a kind of TLS algorithm based on reweighting iteration. During the three-dimensional coordinate conversion we can deal with the observations of gross errors effectively, and it is proven through example that the TLS algorithm based on reweighting iteration can get a more accurate three-dimensional coordinate conversion parameter.

关 键 词：最小二乘(LS) 总体最小二乘(TLS) 选权迭代 转换参数 

分 类 号：P226[天文地球—大地测量学与测量工程]