FFT-PCA Method For Fusing Remote Sensing Imagery

In order to use the combination of spectral and spatial information, the fusion of satellite images are used. The fusion result is an image which includes spectral information of multi-spectral image and spatial information of panchromatic image. This paper investigates the capability of Fast Fourier Transform-Principal Component Analysis (FFT-PCA) method in the fusion of two set of images, including Hyperion and IRS-1D images and IKONOS images, where this method uses the replacement of the panchromatic image with fast Fourier filtering for the purpose of fusion. The fusion results of this method have been compared with the fusion result of Intensity Hue Saturation (IHS), Principal Component Analysis (PCA), Wavelet-Intensity Hue Saturation (Wavelet-IHS), Fast Fourier Transform-Intensity Hue Saturation (FFT-IHS). To compare and analyze the results of the these methods, the criteria for evaluation of the quality of spectral and spatial include correlation coefficient, signal to noise ratio, RMSE, filtered correlation coefficient, SAM and ERGAS were used. The results demonstrate that the FFT-PCA method achieve more precision in image fusion. This method acts more efficient than other methods in terms of information and spectral content preservation of Hyperion and IKONOS images. This method also shows very good performance in preservation of spatial content for IRS