基于PCNN与IFS的可见光与红外图像融合方法

鉴于传统的可见光与红外图像融合方法存在的边缘 模糊与清晰度低等问题,提出了一种基于脉冲耦合神经网络(PCNN)与直觉模糊集(IFS) 的可见光与红外图像融合改进算法。该方法首先利用IHS变换,分离可见光图像的亮度信息 I；其次,利用非下采样轮廓波变化(NSCT)将I分量与红外图像分别进行分解,得到高低频系数；对低频部分采用高斯隶属函数 和直觉模糊集进行融合,对高频部分采用PCNN模型进行融合；再次,通过非下采样轮廓波逆 变化得到融合图像的I分量；最后,进行IHS逆变换得到彩色融合 图像。大量仿真结果表明,这种融合方法能很好地保留可见光与红外光源图像的特征信息和 细节信息,融合后的图像的轮廓更加清晰,具有更良好的视觉效果。与现有的其它红外光和 可见光图像融合方法相比,本文提出的方法,其融合图像的熵值、边缘保持度、互信息、标 准差、结构相似度等指标都有明显的提高,有效地验证了本文算法的有效性。

Research on fusion method of visible and infrared image based on PCNN and IFS

In view of the low edge blurring and l ow definition of traditional image fusion methods,an improved algorithm for visi ble and visible image fusion based on Pulse coupled neural network (PCNN) and in tuitionistic fuzzy set (IFS) is proposed.In this method,the luminance and color information of visible images were separated by IHS,and the luminance I component value was obtained.Then,the I component and the infrared image were decomposed by non-down-sampling contour wave change (NSCT),and the high and low frequency coefficients were obtained,gau ssian membership function and intuitionistic fuzzy set were used to fuse the low frequency parts,and PCNN model was used to fuse the high frequency parts,the I component of the final fusion image was obtained through the inverse change of t he non-lower and non-lower sampled contour waves,and finally the color image w as obtained by the contravariant IHS transformation.Experimental results show th at the final fusion image has more detailed information and the fusion effect is also significantly improved,indicating the superiority of the algorithm.Compare d with other existing infrared and visible image fusion methods,the proposed met hod has significantly improved entropy,edge retention,mutual information,standar d deviation,structural similarity and other indexes of the fusion image,which ef fectively verifies the effectiveness of the proposed algorithm.