基于双树复小波变换的图像融合方法  

Method of Image Fusion Based on Dual-tree Complex Wavelet Transform

作　　者：王亚杰[1] 李殿起[1] 徐心和[1] 

机构地区：[1]东北大学信息科学与工程学院

出　　处：《计算机工程》2008年第15期176-178,共3页Computer Engineering

基　　金：国家自然科学基金资助项目(60475036)

摘　　要：为获得更好的融合效果,提出基于双树复小波变换的图像融合方法。双树复小波变换具有平移不变性、方向选择性等特点,适合进行图像融合,优于传统离散小波变换方法。给出多策略的融合规则,源图像小波变换后低频采用区域清晰度,高频采用区域标准差。灰度多聚焦图像和彩色多聚焦图像的融合实验测试以及评价指标的统计结果,表明了双树复小波变换方法的优势和所用融合规则的有效性。An image fusion method based on Dual-Tree Complex Wavelet Transform（DT-CWT） is put forward in order to obtain better fusion effect. The DT-CWT is suitable to image fusion processing because of its shift invariance and directional selectivity, It is preferable to traditional Discrete Wavelet Transform（DWT）. The fusion rules based on multi-scheme are also presented. The low frequency part employs local definition, and the high frequency parts employ local standard deviation after wavelet transform of source images. The experimental results of gray multi-focus images fusion and color multi-focus images fusion, and the statistics of evaluating index show that DT-CWT is superior to DWT and the proposed fusion rules are effective.

关 键 词：小波变换 双树复小波变换 图像融合 多策略 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]