线阵CCD卫星影像外方位元素求解的研究

Exterior orientation elements computation of satellite linear array scanner imagery

线阵CCD技术已广泛应用于高分辨率卫星遥感影像,但其外方位元素的强相关性使得法方程出现病态,难以得到稳定的解。本文在基于严密成像模型的基础上,采用SPOT卫星影像对几种常见的去相关方法进行了比较,并在谱修正迭代法的基础上,提出了一种新的去相关方法。根据谱修正迭代法中法矩阵的特征,对初始值乘以一个系数,不仅没有改变原有的无偏性质,而且使其对初值的依赖程度大大的降低。分析与试验结果表明,这种改进是有效的。

Linear array CCD has been used in high resolution satellite images widely, but the strong relativity of exterior orientation makes the equation morbid. It is difficult to get the steady solutions. This paper analyzed previous methods for solving the problem of computing the exterior orientation elements based on the strict imaging model. Then it brought forward a better method to avoid the relativity. The method not only kept the agonic characteristic, but also reduced the dependent degree on initial value. Experiments proved that this method was very simple, steady and effective.