第18卷第4期 2010年4月
文章编号
1004-924X(2010)04-0995-07
光学精密工程
Optics and Precision Engineering
利用脉冲耦合神经网络的图像融合
陈浩1,2朱娟1,2,刘艳滢1，王延杰
（1.中国科学院长春光学精密机械与物理研究所，吉林长春130033；
2.中国科学院研究生院，北京100039）
Vol.18No.4
Apr.2010
方法。将多源传感器图像配准后的各个源图像用9/7小波变换的提升算法进行分解，从而得到各个源图像的低频分量和高频分量。对于低频分量，采用像索绝对值选大法进行融合，面高额分量则作为PCNN的输人，在选代结束后，通过比较PCNN点火次数得到一系列融合子图像；然后，用9/7小波的提升算法将获取的一系列多尺度融合子图像进行反变换得到最终的融合图像。设计了可见光图像与红外图像的融合实验，对融合图像的摘、平均梯度、标准差、空间频率进行了定量比较。当使用标准源图像进行融合时，各值比使用传统小波变换与PCNN相结合的图像融合方法分别高 0. 010 4,0. 245 9,0. 113 1 和 0. 284 6,
关键调：红外图像；图像融合；9/7小波；提升算法；脉冲耦合神经网络
中图分类号：TP391.4
文献标识码：A
Imagefusionbasedonpulsecoupledneuralnetwork
CHENHaol-2,ZHUJuan-,LIUYan-ying'，WANGYan-jie(1.ChangchunInstituteof Optics，FineMechanicsandPhysics， ChineseAcademy of Sciences,Changchun130033,China;
2.GraduateUniversity of ChineseAcademy of Sciences,Beijing100039,China)
Abstract: In order to represent a scene exactly and entirely, an image fusion method based on Pulse Coupled Neural Network (PCNN) is proposed. After registering the images of multi-source sensors, obtained images are decomposed into several coefficients of low frequency and high frequency by using the 9/7 wavelet transform based on lifting scheme. The larger absolute gray values are selected to fuse low frequency images and the high frequency images are input to the PCNN, then a serial of fused sub images can be obtained by comparing firing times after the iteration. Finally,the fused images are ob-tained by inversing transform using the 9/7 wavelet based on lifting scheme. By means of design of simulation experiments using visible and infrared images, the entropy, average gradient, standard de-viation and space frequency are selected to evaluate the fused image. Obtained results show that the entropy, average gradient, standard deviation and space frequency of the fused image by using the no vel fusion method base on PCNN are higher 0.010 4, 0.245 9, 0.113 1 and 0.284 6, respectively,
收稿日期：2009-04-23；修订日期：2009-07-23.
基金项目：国家863高技术研究发展计划资助项目（No.2006AA703405F）