Helmert方差分量估计方法已被广泛用于测量平差观测量的定权中,但是,实际应用中该方法却存在不收敛的现象。为此,将信息熵和变异系数引入测量平差,提出了2种新的处理不同类型观测量、同类型不同精度观测值、甚至同类型同精度观测值进行定权的方法。数值实验表明,提出的方法效果优于Helmert方法。

Helmert algorithm has been extensively utilized to estimate the variance components from which the weight of the observations could be identified in a surveying adjustment system.However,the non-convergence of Helmert algorithm has been found in some reality applications.Therefore,we suggested to using entropy theory and variation coefficient to estimate the weight coefficients for different types of observations,the same type of observations with different or same observation accuracy.Numerical experiments made in this study verified that the proposed approach outperforms the Helmert algorithm to provide improved adjustment results.