基于Curvelet变换和小波变换相结合的图像融合算法研究  

Research on Multifocus Image Fusion Algorithm by Combining Curvelet Transform and Wavelet Transform

作　　者：蒋婷婷[1] 王敬东[1] 李鹏[1] 

机构地区：[1]南京航空航天大学自动化学院,南京210016

出　　处：《光电子技术》2010年第2期111-116,共6页Optoelectronic Technology

摘　　要：针对小波变换在图像边缘表达方面的局限性,以及Curvelet变换在表达图像点特征上的不足,提出了在红外图像与可见光图像融合的过程中采用基于Curvelet变换和小波变换相结合的图像融合算法。首先对图像进行Curvelet分解,对低频系数使用基于小波变换的融合算法,对高频系数结合融合图像的特点分别采用了两种不同的选取方法:模值绝对值取大法和基于系数相关性法。最后,对最终系数进行反Curvelet变换,得到融合结果图。采用该算法进行了大量的红外图像与可见光图像融合实验,实验结果表明,此算法的融合结果图获得了更好的目标信息和光谱信息。In order to overcome the weakness of expression of image edge in wavelet transform and feature points in Curvelet transform, a multifocus image fusion algorithm is proposed, which is based on combination of wavelet and Curvelet transform. Firstly, each of the images is decomposed using Curvelet trasnform, among which the low-frequency coefficients are fused by wavelet-based method, while two different select methods are used in high-frequency ones considering feathers of the image：one method is to select a larger value coefficient, and the other one is based on correlation of the two coefficients. Finally, the fused image is reconstructed by performing the inverse Curvelet transform. A large amount of fusion experiments of infrared images and visible images are carried out by the algorithm. Experiment results indicate that better goal and spectral information are obtained.

关 键 词：图像融合 CURVELET变换 小波变换 红外图像 可见光图像 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]