摘要
给出了解决粗差问题的两种计算方案:①将粗差作为待估参数,采用拟稳平差思路解秩亏问题,直接获得粗差;②选取部分观测值作为准观测值,采用部分最小二乘法获得待估参数,将非准观测值的残差作为粗差。结果表明,两种方案与“拟准检定法”具有相同的效果,具有粗差的观测值在平差时不起作用。
Two approaches to gross error detection are proposed in this paper. The first approach is that the gross errors （G） are treated as parameters to be estimated, and the idea of “Quasi-Stable adjust- ment” （Gr^TGr=min）for solving rank deficient problem is adopted. The second approach is that part observations are selected as accurate observations, the principle of partial least-square estimation （Vr^TVr = rain） is used, and than the residuals of the other observations are treated as gross errors. The results by these two approaches are the same as by QUAD（Quasi Accurate Detection）, and the observations with gross error no longer affects the adjustment.
出处
《大地测量与地球动力学》
CSCD
北大核心
2005年第3期29-33,共5页
Journal of Geodesy and Geodynamics
基金
中国科学院动力大地测量学重点实验室开放基金(L04-07)
中国科学院"百人计划"项目
"基础地理信息与数字化技术"
山东省重点开放实验室课题(SD040203)
关键词
平差因子
粗差
准观测值
部分最小二乘法
拟准检定法
adjustment factor, gross error, quasi-accurate observation, part least-square estimation,QUAD（Quasi Accurate Detection）
