基于WBCT与归一化相关矩阵的红外与可见光图像融合算法  

FUSION ALGORITHM OF INFRARED AND VISIBLE IMAGE BASED ON WBCT AND NORMALISED CORRELATION MATRIX

作　　者：苏伍各[1] 陶忠祥[1] 董博[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：《计算机应用与软件》2011年第9期248-250,254,共4页Computer Applications and Software

摘　　要：突出目标指示特性和提高空间分辨率是红外与可见光图像融合的重要问题。针对传统区域能量融合规则的局限性,提出了一种基于小波-Contourlet变换WBCT(Wavelet-based contourlet transform)和归一化相关矩阵与区域能量相结合的融合算法。采用一组红外与可见光图像数据进行融合实验,利用均值、标准差、熵、总体交叉熵和平均互信息评价指标对融合效果进行分析。对结果的分析比较表明,所提出的融合规则的效果优于小波融合方法及Contourlet变换方法,达到了良好的视觉效果。Highlighting target indication characteristic and improving spatial resolution are the important issues to infrared and visible image fusion.Aiming at the limitation of traditional local energy fusion rules,a fusion algorithm is proposed which is based on wavelet-based Contourlet transform（WBCT） and the combination of normalised correlation matrix and local energy.A set of infrared and visual image data are used to carry out the fusion experiment,and five assessment indices are used to analyse the effect of fusion,including mean,standard deviation,entropy,general cross-entropy and average mutual information.Analytical comparison on the results indicate that the effect of the fusion rule proposed in the paper outperforms that of the wavelet fusion method and the Contourlet transform method,and achieves good visual effect.

关 键 词：图像融合 小波-CONTOURLET变换 归一化相关矩阵 结构相似度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]