中文摘要：

在线阵CCD影像外方位元素的求解过程中，外方位元素之间的相关性，导致法方程病态，最小二乘估计值误差较大。本文在分析以往解决方法的基础上，提出采用两步解法求解线阵CCD影像的外方位元素。该算法的实质是选择一个比单位矩阵更合适的正则化矩阵。试验结果表明该算法能有效地克服线阵CCD影像外方位元素间的相关性，定位精度较高，定向点精度在1个像素内，检查点精度在1．5个像素内。

英文摘要：

In computing the exterior orientation elements of array-scanned image, because of the interrelationshi Pof exterior orientation elements, the normal equation is ill-conditioned and the error of least-square estimation is very large. This paper first analyses the previous methods for solving the problem, then puts forward a new method that using two-step solution to compute the exterior orientation elements of array-scanned image. The essential of this new algorithm is to choose a more appropriate regularization matrix than unit matrix. The result of experiment indicates that the arithmetic can effectively overcome the interrelationship among exterior orientation elements and achieve a higher accuracy. The positioning precision can be within one pixel for orientation points and within one and a half pixels for check points.