在测量平差中,如何准确地找出对设计阵的复共线性产生很大影响的观测值(这种观测值称为复共线性强影响观测值),并进行恰当的处理,以便取得高精度的平差成果是一个具有实用价值但目前尚未得到重视的问题。为此,文中回顾和分析了现有的探测方法,在此基础上采用矩阵扰动分析的理论和方法.通过计算每个观测值对条件数的影响率,提出了探测复共线性强影响观测值的新方法——影响比法,并通过算俐验证了该方法的有效性。
In surveying adjustment, how to find.and appropriately deal with the observations named as muhicollinearity influential observations, which may tremendously affect the multicollinearity relations existing among the columns of the design matrix, is valuable and is not focused on. The main objective of this paper is to de velop a diagnostic tool based on the theory of matrix disturbance for the detection of multicollinearity influential observations. A new method named as influence ratio is proposed. One example is given to illustrate its use.