针对有理函数模型具有与传感器类型无关,是一种通用成像模型等特点,以上海地区的QuickBird异轨立体像对为例,利用原始影像提供的有理多项式系数,探讨了基于有理函数模型的卫星影像立体定位算法,系统分析了其定位误差的分布特点.在此基础上,提出了提高基于有理函数模型立体定位精度的多种补偿模型.实验结果表明,在该区域使用一阶仿射变换模型进行补偿,在5~6个均匀分布的控制点支持下,即可达到0.6~0.8m的平面精度和1.0m左右的高程精度.
Rational Function Model (RFM) is one of the universal imaging models with characteristic of sensor independency. The geo-positioning solution based on Rational Polynomial Coefficients (RPC) is studied using QuickBird separate-orbit stereo image pair in Shanghai district, China. The principle of RFM-based stereo positioning method is first introduced, and the positioning errors are then analyzed. Based on this, enhancing models are further presented for the improvement of geo-positioning accuracy. The experimental results show that the affine model works best in the test area. The accuracy can reach to 0.6 - 0.8 m in planimetry and about 1.0 m in elevation based on the affine model by using 5-6 well distributed control points.