【摘要】 图像融合方法同时保持全色图像空间细节和多光谱图像的光谱信息是一个挑战,特别是对于全色和多光谱图像光谱范围部分重合或不重合的波段,或者是波段数较多的多光谱图像,一些融合算法不能较好的支持。针对"资源一号"(ZY-1)02D卫星的全色图像和8波段多光谱图像,对结合光谱响应函数和全局方差匹配的遥感图像融合方法做了适度扩展,并将扩展后的融合方法首次用于ZY-1 02D卫星多个区域的图像融合,并与商业软件使用的效果较好的主流融合方法的结果进行了定性和定量的比较与评价,结果表明:该融合方法空间细节和光谱保持都较好,可用于ZY-1 02D卫星多波段图像融合。更多还原

【Abstract】 It is a challenge to keep both the spatial details of panchromatic images and spectral information of multispectral images, especially for the bands with partial or inconsistent spectral ranges of panchromatic and multispectral images. In addition, some fusion algorithms cannot better support multispectral images with a large number of bands. Aiming at the 8-band multispectral and panchromatic images of ZY-102 D satellite, the remote sensing image fusion method combining the spectral response function and global variance matching is moderately extended. The extended fusion method is applied to the image fusion of multiple regions of ZY-1 02 D satellite for the first time. The method is compared and evaluated qualitatively and quantitatively with mainstream fusion methods used by commercial softwares. The results show that the fusion method has good spatial details and spectral preservation, and can be used for multi-band image fusion of ZY-1 02 D satellite.更多还原