关键词: 超复数, 超复数偶对分解, 多光谱图像, 全色图像, 图像融合, 图像处理

Abstract:

A residual error hypercomplex symplectic decomposition approach to multispectral (MS) and panchromatic (PAN) image fusion is presented. The residual errors of MS and PAN images are modeled by hypercomplex respectively. Hypercomplex symplectic decomposition is applied to the hypercomplex residual error of the MS images along the gray axis of its color space. The result would be the simplex part containing the luminance information of the image and the perplex part including the chrominance information of the image. Analysis indicates that the residual error of high resolution MS images can be recovered once the simplex part of low resolution MS images is replaced by the residual error of high resolution PAN images.Consequently, the fusion of MS and PAN images can be implemented. Simulation results not only verify the effectiveness of the proposed method but also show that there is no distortion of visible spectrums in images fused by this method. Existing performance evaluations for image fusion also demonstrate that the fusion results of this approach are better than those of intensity hue saturation(IHS),principal component analysis (PCA) and wavelet-based fusion methods.

Key words: hypercomplex, hypercomplex symplectic decomposition, multispectral image, panchromatic image, image fusion, image processing