一种基于小波分析的自适应图像融合方法  

AN ADPTIVE IMAGE FUSION ALGORITHM BASED ON WAVELET TRANSFORM

作　　者：柴艳妹[1] 任金昌[1] 赵荣椿[1] 

机构地区：[1]西北工业大学计算机科学与工程系,西安710072

出　　处：《中国体视学与图像分析》2002年第4期240-243,共4页Chinese Journal of Stereology and Image Analysis

基　　金：本文获航空基金支持(编号01F53036)

摘　　要：本文提出了一种基于小波分析的自适应图像融合方法。首先对原始图像进行小波分解,得到四个子图像。然后分别计算各子图像的区域边界强度,再根据不同图像的相应区域边界强度值赋予其不同的加权系数,从而实现自适应的信息融合。最后通过小波反变换,得到完整的融合结果。通过对SAR和TM图像的融合实验,证明了本方法在保持边界特性方面具有明显优势。An adaptive image fusion algorithm based on wavelet transform (WT) is proposed in this paper. Firstly, we decompose the original images into four sub-images via WT. Then, calculate the regional edge intensities (EI) of each sub-image. The corresponded regional EI of different images are further used for determining the weight in adaptive fusion, where we give more weight for the image with bigger EI. Finally, the whole fusion image can be obtained by inverse wavelet transform (IWT). The fusion experiments using SAR and TM images are also done. The experimental results show that the proposed method seems better in preserving edge information.

关 键 词：小波变换 自适应图像融合 边界强度 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]