A Fusion Algorithm of Remote Sensing Images Based on Normalized Correlation Moment of Characteristic Values

Now multi-spectral images and panchromatic images are widely used in remote sensing. In this paper we proposed an algorithm using normalized correlation moment of coefficient characteristic values based on principal component analysis and wavelet transform. The basic idea of this algorithm is to transform the multi-spectral image into three principal components at first, then perform wavelet decomposition of the first component image and the panchromatic image separately. Using normalized correlation moment fusion formula to select the wavelet coefficient of high frequency component in different resolutions. Finally, by performing inverse PCA transform to get the fused image. We compared the fused image using this algorithm with the results of PCA, DWT and other familiar methods, computer simulation shows that the fused image using this algorithm obtain better results in terms of both preserving spectral information and improving spatial resolution of the original multi-spectral images than the others.