摘要
提出了一种基于离散剪切波(shearlet)和改进的脉冲耦合神经网络(PCNN)进行图像融合的方法。首先,采用shearlet变换将已配准的两个源图像进行分解,得到低频子带系数和不同尺度不同方向的高频子带系数,低频部分采用简单的加权平均法;高频部分,提出基于改进的拉普拉斯能量作为PCNN链接强度的算法。最后,进行shearlet反变换得到融合图像。仿真结果表明,本文的算法有更好的融合效果,并且所用时间也比非采样轮廓波(NSCT)少。
An image fusion algorithm is proposed based on shearlet transform and improved Pulse Coupled Neural Networks（PCNN）.Firstly,two registered original images are decomposed by shearlet transform,thus the low frequency subband coefficients and high frequency subband coefficients can be obtained.Secondly,the fusion principle of the low frequency subband coefficients is based on the traditional method of weighted average.As for the high frequency subband coefficients,we present an algorithm which employ the improved laplacian energy as the link intensity of PCNN.The experimental results show that the proposed algorithm outperforms Nonsubsampled Contourlet Transform（NSCT）,and it takes less time than NSCT.
出处
《激光与红外》
CAS
CSCD
北大核心
2012年第2期213-216,共4页
Laser & Infrared
