A Maximum Likelihood Method for Remotely Sensed Image Fusion Using Generalized IHS Transform

Abstract

Image fusion is an effective approach for improving spatial resolution of multispectral(MS) images by incorporating the detailed information of panchromatic (Pan) image, which need not change the original imaging system hardware. In this paper, we propose a fusion method for IKONOS/QuickBird images using generalized IHS (GIHS) technique based on maximum likelihood (ML) estimation. First, the regression coefficients are computed by the regression function between Pan and MS images. Then, the intensity component of the MS images is obtained by those coefficients. After that, a new Pan image is acquired using ML framework based on the observed model. Finally, the fused images are obtained by the GIHS method. Experiments on spatially degraded IKONOS/QuickBird MS images, whose original MS images are available for reference, show that the reconstructed results applying to the proposed approach have better spectral quality and spatial quality.