线阵推扫式影像外定向的一种新算法

A New Method for Linear Pushbroom Imagery Exterior Orientation

线阵推扫式影像外方位元素间的强相关性,导致法方程病态,外方位元素的最小二乘估计值误差较大。本文提出采用岭 压缩组合估计解决线阵列推扫式影像的外定向问题。该算法是综合岭估计和压缩估计的一种新有偏算法,它通过对最小二乘估计值的不同分量进行不同比例的压缩使估计值最优。文中采用10m分辨率的SPOT1全色影像和2.5m分辨率的SPOT5全色影像进行实验,并与商业遥感软件Erdas的实验结果进行比较。研究结果表明该算法能有效地克服线阵推扫式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,达到Erdas软件的定位精度。

The strong interrelationship among exterior elements of linear pushbroom imagery induces normal equation severely ill-posed and least squares value deviated from the true value. In this paper the Ridge-Stein combined estimator (RSC) is presented for linear pushbroom imagery exterior orientation. The estimator is a new biased method combining the Ridge estimator and the Stein estimator. It achieves optimum estimation values through applying different scale compression to each least squares component. In the paper we make experiments with one 10-meter SPOT 1 panchromatic image and one 2.5-meter SPOT 5 panchromatic image, and compare the results with values computed using commercial remote-sensing software Erdas. The research results show that the algorithm can effectively overcome the strong interrelationship among exterior elements and achieve high accuracy. The position precision can be within one pixel for orientation points and within one and a half pixels for check points, as accurate as that of Erdas software.