基于无下采样小波-Contourlet变换的图像融合算法  

Image Fusion Based on Nonsubsampled Wavelet-contourlet Transform

作　　者：王连庆[1] 安锦文[1] 

机构地区：[1]西北工业大学自动化学院,陕西西安710072

出　　处：《激光与红外》2008年第2期191-194,共4页Laser & Infrared

摘　　要：针对目前最新发展的无下采样Contourlet变换,同时具有方向性、各向异性和平移不变性的特点,并结合多孔小波变换的优点,提出了一种新的基于无下采样小波-Contourlet变换图像融合算法。通过对航空图像进行无下采样小波-Contourlet变换的多分辨分解得到高频和低频系数,并根据不同分解面的系数特性,采用不同的加权融合规则进行融合。实验结果表明,此方法相对于传统的变换法,取得了更佳的融合结果。A novel method of image fusion based on the nonsubsampled wavelet- contourlet transform is proposed, combined with the merit of a trous transform, based on the recently developed nonsubsampled contourlet transform, which can offer a trait of flexible multiscale and shift-invariant simultaneity. The high and low frequency-coefficients are obtained by decompose the aeronautic image with the nonsubsampled wavelet-contourlet transformation. Then, according to the different coefficient characters, these coefficients are syncretized by using different fusion algorithms of weighted average. The experimental results show that the new method makes a better fusion result compared with the traditional transform methods.

关 键 词：图像融合 无下采样Contourlet交换 多孔小波变换 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]