改进的双通道自适应PCNN和NSST的遥感图像融合  

Non-Subsampling Shearlet Transform Remote Sensing Image Fusion with Improved Dual-channel Adaptive Pulse Coupled Neural Network

作　　者：阮理念 董燕[1] Ruan Linian;Dong Yan(Faculty of Land and Resources Engineering,Kunming University of Science and Technology,Kunming 650032,Yunnan,China)

机构地区：[1]昆明理工大学国土资源工程学院,云南昆明650032

出　　处：《激光与光电子学进展》2023年第10期364-374,共11页Laser & Optoelectronics Progress

摘　　要：遥感图像融合作为一种整合多光谱和全色图像所包含信息的有效方法,在国土空间规划和灾情检测等应用领域已成为一种强大的技术。针对非下采样剪切波变换(NSST)域的融合策略进行研究,提出一种新的NSST域遥感图像融合方法。首先对源图像进行NSST,将其分解为低频系数和多方向的高频子带;然后,引入基于平均谱半径(MSR)的图像特征加权机制,将能量属性和改进的拉普拉斯能量和进行加权并应用于低频系数融合,以解决能量保存和细节提取问题;其次,开发一种改进的双通道脉冲耦合神经网络,并结合由方向信息确定权重的加权自适应方法来对高频子带进行融合;最后,利用融合后的低频系数和高频子带进行重构,得到融合后的图像。通过GF-2、GeoEye和WorldView-3这3种不同分辨率的共48组卫星影像验证了该方法的有效性。与5种融合方法的对比实验表明,该方法在视觉感知和定量评价指标方面都能取得较好的效果。Remote sensing image fusion,an effective method that integrates information contained in multispectral and panchromatic images,has become a powerful application technology in fields such as territorial spatial planning and disaster detection.A new method of remote sensing image fusion in nonsubsampling shearlet transform(NSST)domain is proposed on the basis of research on fusion strategy in the NSST domain.First,the source image is NSST decomposed into lowfrequency coefficients and multidirectional highfrequency subbands.Subsequently,to solve the problems of energy conservation and detail extraction,an image feature weighting mechanism based on the mean spectral radius weights the energy attribute and the improved Laplacian energy sum and applies these to lowfrequency coefficient fusion.Next,an improved dualchannel pulse coupled neural network is developed to fuse the highfrequency subbands by combining the weighted adaptive method with the direction information to determine the weight.Finally,the fused lowfrequency coefficients and highfrequency subbands are used for reconstruction to obtain the fused image.The effectiveness of this method is verified by 48 sets of satellite images with three different resolutions,namely,GF2,GeoEye,and WorldView3.The comparison experiment with five fusion methods shows this new method can achieve good results in visual perception and quantitative evaluation indicators.

关 键 词：遥感 图像融合 非下采样剪切波变换 双通道脉冲耦合神经网络 加权自适应 平均谱半径 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]