基于NSCT变换和PCA的图像融合算法  

Image Fusion Algorithm Based on NSCT and PCA

作　　者：吴粉侠[1] 李红[1] 李洪星[1] 

机构地区：[1]咸阳师范学院图形图像研究所,陕西咸阳712000

出　　处：《航空计算技术》2015年第3期47-51,共5页Aeronautical Computing Technique

基　　金：陕西省科技厅专项科研计划项目资助(2013JM8035);国家级创新训练项目资助(20131072002);咸阳师范学院科研基金项目资助(13XSYK058)

摘　　要：针对小波变换与Contourlet变换的不足,提出了一种将非下采样Contourlet变换与主成分分析相结合的图像融合算法。对已配准好的融合源图像作NSCT分解,采用PCA加权融合规则对低频系数处理,采用边缘强度最大的规则对各高频子带系数作融合处理,分别得到融合图像的变换系数,对该系数经NSCT逆变换重构得到融合图像。实验结果表明,算法有效地融合了源图像中的重要信息,融合图像边缘、纹理、细节清晰,得到较好视觉效果和较优的评价指标。In this paper,a novel algorithm is proposed combining Nonsubsampled Contourlet Transform and Principal Component Analysis to avoid the disadvantages of Wavelet transform and Contourlet transform. The registered images are decomposed by NSCT. We can obtain the low- frequency subimages and a series high- frequency subimages. The weighted PCA is applied to the low- frequency subimages. And the rule of edge intensity is applied to the high- frequency subimages. Then the fusion image is obtained by inverse NSCT. Experimental results show that the proposed method can retain a visual quality and objective evaluation index,and performs some related fusion approaches.

关 键 词：图像融合 非采样CONTOURLET变换 边缘强度 主成分分析 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]