利用GA求解卫星影像的空间后方交会

Spatial Resection of Satellite Images with a Genetic Algorithm

克服经典平差线性化的不足,将遗传算法理论引入卫星影像的空间后方交会解算中,利用遗传算法全局和局部搜索力强的优势,求解IKONOS影像和模拟影像的单片空间后方交会病态问题。实验结果表明,遗传算法可以有效求得精度较高的最优解;同时,相比最小二乘、岭估计等方法,其运行效率也较高。

A Genetic algorithm is used to overcome the disadvantages of classical least squares and estimate nonlinear least squares, this algorithm has the advantage of global and local search capabilities. In this paper, genetic algorithm theory is introduced to the spatial resec- tion of IKONOS and simulation images. Experiments show that a genetic algorithm can solve the ill-conditioned equations of spatial resection and achieve optimal solutions with higher precision with greater operating efficiency than other methods.