基于非子采样Contourlet变换的多聚焦图像融合方法  

Multi-focus Image Fusion Algorithm Based on Nonsubsampled Contourlet Transform

作　　者：邹蓉[1] 李玲玲[1] 李翠华[1] 

机构地区：[1]厦门大学信息科学与技术学院,福建厦门361005

出　　处：《厦门大学学报（自然科学版）》2009年第6期812-817,共6页Journal of Xiamen University：Natural Science

基　　金：国家重点基础研究发展计划(2007CB311005);河南省基础与前沿技术研究计划(092300410043);河南省高等学校青年骨干教师资助计划;河南省教育厅自然科学基础研究计划(2007510023);福建省自然科学基金计划资助(A0710020);厦门大学985二期信息创新平台资助

摘　　要：非子采样Contourlet变换(NSCT),是针对Contourlet变换的一些局限性进行了改进,可以对图像进行灵活的多尺度、多方向和平移不变性分解.提出了一种基于NSCT的多聚焦图像融合方法.首先对多聚焦图像进行NSCT变换;然后对变换得到的低频分量系数采用改进的加权平均融合规则进行融合处理,对高频分量的最高层和其它层系数分别采用绝对值最大和改进的区域方差融合规则进行融合处理;最后重构图像得到融合结果;并给出了实验结果.对结果的分析比较表明,所提出的融合规则的效果优于常用的融合方法和参考文献的融合方法.Nonsubsampled contourlet trsnsform（NSCT） is an improvement based on contourlet transform. It is a flexible multistage, multidirectional and shift invariant. In this paper, an image fusion method based on NSCT is proposed. Firstly images are decomposed by NSCT. Then a low frequency fusion rule based on the improved weighted average is used to fuse the low frequency bands, while the fusion rule with maximum absolute value coefficient and based on the improved regional variance is used to fuse the high frequency bands. Finally, the fused coefficients are reconstructed to obtain fusion results. The experimental results are given, and compared with those of the common fusion methods and some referencest fusion methods. The experiments show that the approach proposed in this paper can achieve better results than other methods.

关 键 词：图像融合 无下采样Contourlet变换(NSCT) CONTOURLET变换 多聚焦图像 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]