摘要
为了克服小波变换与Contourlet变换在图像融合应用中的不足,文中提出了一种基于非下采样Contourlet变换(Non-Subsampled Contourlet Transform,NSCT)与主成分分析(Principal Component Analysis)及脉冲耦合神经网络(PulseCoupled Neural Networks,PCNN)相结合的融合算法。首先对被融合图像执行NSCT分解,将图像分解为一个低频子带图像和多个高频子带图像,利用PCA加权规则融合低频子带图像得到融合图像的低频子带图像,利用PCNN规则融合各高频子带图像得到融合图像的高频子带图像,最后经NSCT逆变换重构图像。实验结果表明,该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,得到了较好的视觉效果和较优的评价指标。
Toovercome the shortage of wavelet transformation and Contourlet transformation in image fusion, a novel algorithm is pro- posed combining Non-Subsampled Contourlet Transform （NSCT）, Principal Component Analysis （PCA） and Pulse-Coupled Neural Networks （PCNN）. The registered images are decomposed by NSCT, which can obtain the low-frequency subimages and a series high- frequency subimages. The rule of PCA is applied to the low-frequency subimage. And the rule of PCNN 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.
出处
《计算机技术与发展》
2015年第12期72-75,79,共5页
Computer Technology and Development
基金
国家级创新训练项目(20131072002)
陕西省教育科研项目(14JK1802)
咸阳师范学院科研基金项目(13XSYK058)
