一种基于NSCT变换的红外与可见光图像融合算法  

A Kind of Fusion Algorithm for Infrared and Visible Images Based on NSCT

作　　者：王跃华[1] 陶忠祥[1] 

机构地区：[1]空军航空大学航空军械工程系,吉林长春130022

出　　处：《光学与光电技术》2012年第4期81-84,共4页Optics & Optoelectronic Technology

摘　　要：针对红外与可见光图像融合,提出了一种基于NSCT变换的图像融合方法。对经NSCT变换的低频子带系数采用基于区域能量自适应加权的融合规则,对高频子带系数采用混合的融合方法,即对于低层,采用基于区域方差选大的融合方法,对于高层采用像素点的绝对值选大的融合方法。实验结果表明,该融合算法可以获得更多的细节信息,能获得较理想的融合图像。Aiming at the fusion of infrared and visible images, a novel fusion algorithm based on Nonsubsampled Contourlet Transform（NSCT） is proposed. After the Nonsubsampled Contourlet Transform of the source images, different fusion rules are employed to deal with different image informations. The low-frequency approximation image is processed by using adap rive regional energy, and the high-frequency detail images in different levels are processed by different fusion rules. The high yield high-frequency detail information is fused by using absolute maximum based on detail coefficients and the high yield high-frequency detail information is fused by using maximum regional variance. The experimental results show that the pro posed algorithm can effectively get more detail information, and the fused images have good qualities.

关 键 词：图像融合 NSCT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]