QuickBird全色与多光谱影像融合方法比较研究

Study on Fusion Algorithms of QuickBird PAN and Multi Spectral Images

自从高分辨率卫星影像出现以来,遥感影像的融合一直是一个重要的研究领域。随着遥感影像获取手段的增加和分辨率的提高,影像的融合不仅仅要求增加空间信息而且在原始多光谱波段的光谱保持方面也有越来越高的要求。目前图像融合已研究出多种方法,如主成分变换法(PCA)、乘积法、Brovey变换法、IHS变换法、SVR变换法以及小波变换法等。这些方法在中低分辨率影像融合时效果较好但却不一定适用于像QuickBird、IKONOS等高分辨遥感影像的融合。本文应用Pansharp变换法和Gram-Schmidt变换法两种融合方法,对QuickBird影像进行融合试验,并与Brovey变换法、IHS变换法、PCA变换法的融合效果进行比较,以评价各算法的优缺点。研究结果表明,Gram-Schmidt变换和Pansharp变换在QuickBird影像融合时不仅能增加空间信息而且还能较好地保持原多光谱波段的光谱信息,其中以Gram-Schmidt变换法所获得的效果最好。

Since the advent of high spatial resolution satellite images, the fusion of multi-resolution images has been an important field of research. With the development of quantitative remote sensing, not only improving spatial details but also preserving the spectral information of muhispeetral bands were required. Many methods, such as Principal Component Analysis （PCA）, Multiplicative Transform, Brovey Transform, IHS Transform, SVR Transform and wavelet transform, have been developed. These fusion methods are successful on fusion of low and medium resolution satellite images, but may not be efficient for the new high resolution satellite images, like QuickBird, IKONOS, as well. In this paper, two new fusion methods with high fidelity of spectral information （Gram-Schmidt transform and Pansharp transform） and Brovey Transform, IHS Transform, PCA Transform were applied and compared, and their advantage and disadvantage were discussed. The results showed that the QuiekBird images fused by Gram-Schmidt transform and Pansharp transform, especially the Gram-Schmidt transform, represented almost the same colour as the original muhispectral images and the same detail as the original panchromatic images.