基于加权整体最小二乘的单像空间后方交会解算

The Solution of the Single Image Space Resection Based on Weighted Global Least Squares

在常规单像空间后方交会的解算过程中,整体最小二乘(TLS)算法在减小地面控制点及观测变量存在的误差方面有较广泛的应用。但在解算不等精度测量时,直接采用TLS算法进行参数估计易出现解失真的情况,因此为提高解算精度,可以为不同观测值定权,进行外方位元素的计算。本文以解算全微分共线方程为研究主线,并引进协因数阵为观测值定权,尝试采用加权整体最小二乘算法(WTLS)对单像空间后方交会进行解算。实验结果表明,本文算法较最小二乘算法与整体最小二乘算法在精度和准确度方面均有较大提高,在单像空间后方交会解算中具有较好的应用前景。

The total least squares(TLS)algorithm has a wide application in reducing the error of ground control points and observation variables.However,for the unequal precision measurement,it is easy to get the distorted solutions with the TLS algorithm.Therefore,in order to improve the calculation precision,the weights of different observations can be determined and then the outer-position elements can be calculated.This paper takes the solution of the full differential collinear equation as the main research line,introduces the covariate matrix to determine the weights for the observation values,and tries to use the weighted total least squares algorithm(WTLS)to solve the single image space resection.The results of the experiments show that in comparison of the least squares algorithm and the total least squares algorithm,the WTLS algorithm presented in this paper has significant improvement in precision and accuracy,which has a good application prospect in the solution of the single image space resection.