基于小波变换的区域特征加权自适应图像融合算法  

Weighting of Regional Features Self-Adaptive Image Fusion Method Based on Wavelet Transform

作　　者：郑永安[1] 宋建社[1] 曹继平[1] 刘迎春[1] 

机构地区：[1]西安高技术研究所,西安710025

出　　处：《计算机工程与应用》2004年第6期94-95,127,共3页Computer Engineering and Applications

基　　金：国家自然科学基金重点资助项目(编号:69931040

摘　　要：提出了一种基于小波变换的区域特征加权自适应图像融合算法。首先对源图像做小波变换,分解成低频和高频图像,分解后的低频图像选取一组区域特征进行自适应动态加权,高低频图像采用区域能量融合规则,再对得到的低频和高频图像进行小波反变换。实验结果表明该方法能得到较好的融合效果。In this paper,a weighting of regional features self-adaptive image fusion method based on wavelet transform is proposed.Firstly,the wavelet transform is used to perform a multiscale decomposition of each image.Then,a set of re-gional features are extracted respectively from low-frequency images.To high-frequency images,some special regional en-ergy rules are used.The result images is generated by reverse wavelet transform based on the low and high frequency images.The experiment result appears the proposed method makes obviously effects.

关 键 词：图像融合 小波变换 MALLAT算法 区域特征 自适应融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]