基于EI优化权值的多聚焦图像融合方法  

Multi-focus image fusion method based on weights optimized by EI

作　　者：李婷[1] 谭勇[1] 夏东坤[1] 陈学佺[1] 

机构地区：[1]中国科学技术大学电子工程与信息科学系,合肥230027

出　　处：《计算机工程与应用》2008年第15期192-194,201,共4页Computer Engineering and Applications

基　　金：中国科学技术大学青年基金(the Youth Foundation of the University of Science and Technology of China under Grant No.0610100001)

摘　　要：针对图像融合空间域方法中权值确定的问题,提出了一种基于图像融合客观评价准则中的边缘信息E(IEdge Information)的优化权值方法,通过寻找最优权值使得融合图像的EI最大,从而获得最好的融合效果。该方法采用适应度函数为EI的遗传算法实现,并且结合分块加权融合方法融合多聚焦图像,得到最优融合结果。实验结果表明,该方法有良好的融合效果且优于传统的融合方法。To solve the problem of determining the weights in the spatial domain method of image fusion,a method based on weights optimized by EI （Edge Information） of objective image fusion measure is proposed,which searches the optimized weights to make the biggest EI of the fusion image,then obtains the best fusion results.The method employs generic algorithm and uses the EI as fitness function of genetic algorithm.For the multi-focus images,the optimized fusion image can be obtained by block-weighted fusion method.Experimental results show that the proposed method is better than the traditional methods.

关 键 词：边缘信息 权值 多聚焦图像融合 遗传算法 分块加权融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]