后方交会模型病态问题分析与改善

Analysis and optimization of ill-posed model in space resection

病态性问题一直广泛存在于各种数据处理中。摄影测量学中的许多基本模型在有些条件下也容易使方程组产生病态,如在后方交会时加入自检校解算内方位元素等情况下易得到不稳定的解。主要介绍摄影测量中的一种基本模型——后方交会,以及病态性的基本概念、几何含义、危害性和几种改善算法。实验部分对后方交会中的病态问题进行研究,并利用岭估计及截断奇异值法进行改善分析,提高解算精度。

Ill-posed problem widely exists in data processing. Many models in photogrammetry can eastly cause ill-posed equations under some conditions. For example, adding self-calibration in resection is easy to obtain unstable solutions of orientation elements. It presents a basic model of photogrammetry and space resection. It also presents the basic concept of ill-posed condition, as well as the harmfulness and several algorithms to improve the results. The experiments mainly aim at the ill posed model in resection, with ridge estimate and truncated singular value decomposition method to improve the accuracy.