【摘要】 基于IKONOS-2全色与多光谱图像和Cartosat-1全色与CBERS-02多光谱图像,通过模拟的和真实的配准误差,采用IHS融合和Brovey融合方法,本文讨论了配准误差对融合结果的影响。实验表明,配准误差对融合质量的影响较大,在遥感图像融合处理中,配准误差越小越好;对于保持光谱信息的融合,当配准误差增加时,两种方法产生的光谱变形都迅速增加,此后光谱变形程度随配准误差的变大继续缓慢增强;对于保持空间信息的融合,配准误差的增大,导致两种方法融合结果的清晰度都下降,但下降速度略有不同。更多还原

【Abstract】 This paper mainly discussed the effects of registration residual in IHS and Brovey image fusion methods,using two datasets,IKONOS-2 panchromatic/multispectral images and Cartosat-1 panchromatic/CBERS-02 multispectral images,with simulated and reality registration residuals.Experiments showed that: registration residual affects two methods a lot,and must be limited as small as possible in the applications;for the image fusion of spectral information preserving,spectrum distortion augments heavily while the registration residual occurs,and then increases much more slowly as registration residual becomes larger;for the image fusion of spatial information preserving,definition decreases for both methods with approximate speeds while registration residual increases.更多还原