Abstract: An adaptive fusion algorithm based on Pulse Coupled Neural Networks（PCNN） and Fruit Fly Optimization Algorithm（FOA） is proposed in order to overcome the difficulty of parameters selection of PCNN. The mean structure similarity is used as fitness function of FOA and the global search ability of FOA is used to set four parameters of PCNN. The source images are fused by PCNN with maximum principle. The experimental results demonstrate that the proposed method outperforms the other methods in term of visual evaluation and objective evaluation.

Key words: image fusion, Pulse Coupled Neural Networks（PCNN）, Fruit Fly Optimization Algorithm（FOA）, mean structural similarity

摘要： 针对脉冲耦合神经网络（Pulse Coupled Neural Networks，PCNN）中参数选取不易确定的不足，提出一种基于脉冲耦合神经网络和果蝇优化算法（Fruit Fly Optimization Algorithm，FOA）的自适应图像融合算法。利用FOA的全局搜索能力，以平均结构相似度作为FOA的适应度函数，对PCNN的4个参数[β、][Vθ、][αL]和[αθ]进行自适应设定；结合最大化原则，采用PCNN对源图像进行融合。实验结果表明，该算法在主观视觉效果和客观评价指标上优于其他融合算法。

关键词: 图像融合, 脉冲耦合神经网络, 果蝇优化算法, 平均结构相似度