一种基于多小波变换的方向加权图像融合算法  

AN ALGORITHM OF IMAGE FUSION WITH DIRECTIONAL WEIGHTING BASED ON MULTI-WAVELET TRANSFORM

作　　者：孙湧[1] 孙宏伟[1] 戴光智[1] 

机构地区：[1]深圳职业技术学院电信学院,广东深圳518055

出　　处：《计算机应用与软件》2009年第8期107-109,173,共4页Computer Applications and Software

基　　金：粤港关键领域重点合作项目(2005A11304003)

摘　　要：针对单小波在图像变换中存在的不足,将研究推到多小波领域。提出边沿信息方向性加权的区域能量融合改进算法,重新构造融合图像对应的多小波系数,通过逆变换重构融合图像。此外,利用交叉熵;平均误差;峰值信噪比等参量,对该融合方法的融合性能进行了评价与分析。实验结果表明,该融合方法是有效的。In light of the shortage of single-wavelet in image transform, research subject has been progressed to multi-wavelet field. An improved algorithm of region energy image fusion with directional weighting along the edge information is proposed, which reconstructs wavelet coefficients corresponding to fused images, and then reconstructs fused images through inverse transform. In addition, cross-entropy, average error, peak signal to noise ratio and other parameters are used to evaluate and analyze the fusing property of the fusion method. Experimental resuits show that the fusion method is effective.

关 键 词：多小波变换 图像融合 方向加权 SA4多小波基 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN919.81[自动化与计算机技术—计算机科学与技术]