【摘要】 针对传统脉冲耦合神经网络(PCNN)图像融合算法中最优融合结果无法自适应确定及神经元参数取固定常数所造成的同步脉冲周期无法随图像特征改变的不足,提出了一种基于人工鱼群寻优的自适应双通道PCNN图像融合算法。利用合成空间雷达(SAR)图像的辐射分辨率和可见光图像的清晰度分别作为双通道PCNN对应神经元的链接强度值,PCNN的信号衰减常数、阈值放大系数和水平调节因子3个参数采用人工鱼群寻优,目标函数由互信息(MI)和结构相似度(SSIM)两种图像质量评价指标构建,最终获得近似最优的融合图像。实验结果表明,本文算法图像融合结果优于传统拉普拉斯变换、离散小波变换和参数取固定值的PCNN图像融合算法及其一些改进算法。更多还原

【Abstract】 Aiming at the problem that optimal image fusion result is difficult to be determined adaptively and the parameters of traditional pulse coupled neural network(PCNN)are usually fixed which result in the synchronization pulse period can′t be changed with the change of image feature,the adaptive dual channel PCNN image fusion algorithm based on artificial fish swarm parameter optimization is presented in this paper.The definition of the visible image region and the radiation resolution of the synthetic space radar(SAR)image is used as the link strength of the dual channel PCNN.The parameters including signal attenuation constant,threshold amplification factor and horizontal adjustment factor of the PCNN are determined by artificial fish swarm optimization.Optimization objective function of artificial fish swarm is constructed by using mutual information(MI)and structural similarity(SSIM).Finally,the approximate optimal fusion image can be obtained.Results demonstrate that the proposed algorithm is superior to the Laplacian transform algorithm,discrete wavelet transform algorithm and traditional PCNN algorithm.更多还原