基于Q-Shift DT-CWT的多聚焦图像融合研究  

Multifocus Image Fusion Method Based on Q-Shift DT-CWT

作　　者：徐月美[1,2] 张虹[1] 张凤云[2] 

机构地区：[1]中国矿业大学环测学院,江苏徐州221116 [2]中国矿业大学计算机学院,江苏徐州221116

出　　处：《微电子学与计算机》2011年第9期206-209,共4页Microelectronics & Computer

摘　　要：提出了一种基于Q-Shift双树复小波变换的多聚焦图像融合算法.根据多聚焦图像的成像特点和变换后的高低频系数相关性,对高频系数采用"模值绝对值和取大"和对低频系数采用"局部区域标准方差取大"的融合准则,并对高频融合系数进行一致性检测,以实现尽可能直接选择源图像中的清晰区域系数作为融合图像对应位置的系数.实验结果表明,该融合方法获得了很好的融合效果,与小波变换相比,充分显示了近似的平移不变性和良好方向选择性等特性.The paper proposes a multifocus image fusion method which is based on the Q-Shift dual-tree complex wavelet transform.The low frequency coefficients and high frequency coefficients were fused with local area standard deviation maximum selectivity and module absolute value sum maximum selectivity according to the multifocus imaging features and the correlation of low frequency coefficients and high frequency coefficient.Then it is important to verify consistency of the fused high frequency coefficients.The ideal fused image is to directly use the vivid area coefficients of the source image as the coefficients of the fused image.Experience results show that the paper＇s image fusion method has a better quality,and it shows the approximate shift invariant and good directional selectivity of Q-Shift DT-CWT comparing to the wavelet transform.

关 键 词：Q-Shift双树复小波变换 多聚焦图像融合 模值绝对值和 局部区域标准方差 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]