基于Directionlet变换的多聚焦图像融合  

Multi-focus Image Fusion Algorithm Based on Directionlet Transform

作　　者：顾勇[1] 吕英丽[1] 张晓峰[1] 

机构地区：[1]河北建筑工程学院电气工程系,河北张家口075000

出　　处：《电视技术》2013年第11期51-53,共3页Video Engineering

摘　　要：针对多聚焦图像成像的特点,在同一场景内很难获得全部清晰的图像,在分析Directionlet变换的基础上,提出了一种基于Directionlet变换结合图像区域特征的融合方法,利用Directionlet变换把图像分解成低频和高频子图像,对于变换后的高频子图像,选择区域特征明显的作为融合后的子图像;对于低频子图像,采用加权融合;最后进行Directionlet逆变换,得到融合后的图像,并对融合后图像采用信息熵、平均梯度、相关系数等指标进行评价。实验结果表明该算法能够较好地保留原图像的细节特征信息,具有较好的视觉效果,优于普通小波融合的方法。Because of the characteristics of multi-focus image,it is often difficult to get distinct images in focusing at different depths. After the theory of directionlet transform is put forward, a new muhi-focus image fusion algorithm based on regional feature is proposed in this paper. Firstly, the multi-focus images are decomposed using directionlet transform, which is multi-scale and multi-direction characteristics. Then, the low frequency coefficients are combined weighted evenly, while high frequency coefficients are chosen based on regional features. Finally, the fused image is obtained by performing inverse directionlet transform. And then entropy, average gradient, correlation coefficients are adopted to evaluate the fused image. The result shows that better visual effect can be achieved via this method and the significant information of the original image maintained,which proves that the algorithm is better than the traditional wavelet methods.

关 键 词：Directionlet变换 图像融合 融合规则 区域特征 

分 类 号：TN911.73[电子电信—通信与信息系统] TP391[电子电信—信息与通信工程]