摘要
介绍总体最小二乘的奇异值分解法(SVD)和混合总体最小二乘法(LS-TLS),基于间接平差原理推导一种总体最小二乘迭代解法,可以用来解决系数矩阵含常数列的总体最小二乘平差问题。最后分别对系数矩阵不含常数列和系数矩阵含常数列的算例进行验证,得到的结果与采用奇异值分解法和混合总体最小二乘法计算的结果相同,表明算法的有效性。
It briefly introduces a total least squares singular value decomposition （SVD） and mixed totalleast squares （LS-TLS）. An iterative algorithm for total least squares is derived based on a principle of indirect adjustment, which can be used to solve the adjustment problems of total least squares for coefficient matrix containing constant sequence. Finally, numerical examples of the coefficient matrix without constant sequence and with constant series are analyzed respectively. The results are the same used by the singular value decomposition method and the mixed total least squares and shows the effectiveness ofthe algorithm.
出处
《测绘工程》
CSCD
2014年第7期38-40,45,共4页
Engineering of Surveying and Mapping
关键词
总体最小二乘
间接平差
迭代解法
混合总体最小二乘
奇异值分解
total least squares
indirect adjustment
iteration algorithm
mixed total least squares （LS-TLS）
singular value decomposition （SVD）
