基于双树复小波变换的多聚焦图像融合方法  

Multi-Focus Image Fusion Method Based on Dual-Tree Complex Wavelet Transform

作　　者：杨伟[1,2] 柴奇[1,3] 王黎明[1] 

机构地区：[1]合肥电子工程学院安徽省红外与低温等离子体重点实验室 [2]92602部队 [3]73678部队

出　　处：《光电技术应用》2009年第3期59-62,共4页Electro-Optic Technology Application

基　　金：安徽省重点实验室基金(2007A0103013Y)

摘　　要：针对同一场景的多聚焦图像融合,提出了一种基于双树复小波变换(DT-CWT)的图像融合新算法.首先利用DT-CWT对图像进行多尺度和多方向分解,并根据双树复小波分解域各子带的系数特性定义了图像局部方向对比度,然后针对高频分量系数的选择,采用基于方向对比度的融合规则,而在低频域采用图像清晰度为测度的融合策略.实验结果表明,该算法能够很好地将多聚焦图像中的重要信息提取并注入到融合图像中,与其他方法相比较,取得了更好的融合效果,提高了融合图像的质量.Focusing on multi-focus image fusion of the same scene, a new image fusion algorithm based on dual-tree complex wavelet transform （DT-CWT） was proposed. Firstly the image was decomposed with various scales and directional features using DT-CWT, the local directive contrast was defined according to the charac- ters of the coefficients of all the sub-bands in the domain of dual-tree wavelet transform, and then the local direc- tive contrast was applied in high frequency parts, and the definition of the image was applied in low frequency parts. Compared with other methods, the experiment results show that the proposed algorithm can extract and infuse the important information of the multi-focus image to the fusion image, the better fusion results are ac- quired, and the quality of fusion image was improved.

关 键 词：图像融合 双树复小波变换 清晰度 平移不变性 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]