基于二通道不可分加性小波的多光谱图像融合  

Multi-Spectral Image Fusion Based on Two-Channel Non-Separable Additive Wavelets

作　　者：刘斌[1] 彭嘉雄[2] 

机构地区：[1]湖北大学数学与计算机科学学院,武汉430062 [2]华中科技大学图像识别与人工智能研究所,武汉430074

出　　处：《光学学报》2007年第8期1419-1424,共6页Acta Optica Sinica

基　　金：国家自然科学基金(10477007);湖北省自然科学基金(2006ABA015);湖北省教育厅重点项目基金(D200510004)资助课题

摘　　要：提出了伸缩矩阵为[1,1;1,-1]的不可分小波的一种构造方法,并把它应用于多光谱图像与高分辨力全色图像的融合中。提出了一种二通道对称的不可分小波滤波器组的构造方法,设计出多组具有紧支撑、对称性和正交性的不可分小波6×6滤波器组,利用此类滤波器组中的低通滤波器对图像进行加性分解与重构。采用三种模式NAWS、NAWRGB和NAWL对ETM+及其它类多光谱图像与高分辨力全色图像进行融合研究。实验结果表明,该方法对多光谱图像与高分辨力全色图像的融合有较好的融合效果,与传统的融合方法相比,该方法既能保持多光谱图像的光谱信息,又能保持高分辨力全色图像的高分辨力特性,克服了传统张量积小波融合方法不能得到高分辨力图像的不足,并能节约运算量。A construction method of non-separable wavelets with dilation matrix of [1,1 ; 1, - 1] is proposed and applied to the fusion of multi-spectral images and panchromatic images with high resolution. A construction method of two-channel symmetric non-separable wavelets filter bank is also proposed, and non-separable wavelets 6 × 6 filter banks with compact support, symmetry, orthogonality are designed. The images are discomposed and reconstructed through additive wavelet mode with low-pass filters of the above filter bank. ETM＋ images and other multi-spectral images and panchromatic images with high resolution are fused and studied through NAWS, AWRGB and NAWL wavelet modes. The experimental results show that this method has well fusion performance for multi-spectral images and panchromatic images with high resolution. Compared with other fusion methods, the proposed method preserves the spectral information of multi-spectral images and the high resolution of panchromatic images. It overcomes the shortage of traditional vector product wavelet fusion method, i.e. being incapable of get images with high resolution, and saves computation time.

关 键 词：图像处理 不可分小波 图像融合 滤波器组 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]