摘要
由于传统稀疏表示(SR)冗余字典单一,脉冲耦合神经网络(PCNN)模型参数设置复杂,为了解决上述问题,提出了基于非下采样剪切波变换(NSST)的红外与可见光图像融合算法。该算法首先通过NSST将源图像分解成低频子带和高频子带。然后,使用自适应稀疏表示(ASR)模型进行NSST域低频部分稀疏系数的融合;同时,采用参数自适应脉冲耦合神经网络(PA-PCNN)模型融合相应的高频部分。最后,对融合后的低频和高频波段进行NSST反变换,重建得到融合结果图。实验结果表明:该算法解决了传统SR模型的"块效应"问题,克服了PCNN模型中自由参数的设置难点,在主观视觉和客观评价上均优于现有算法。
Due to the conventional sparse representation(SR) redundant dictionary is single and the parameter setting of pulse-coupled neural network(PCNN) model is complex.To solve the above problems,an infrared and visible image fusion algorithm based on the non-subsampled shearlet transform(NSST) is proposed.Firstly,the algorithm decomposes the source image into low-frequency subband and high-frequency subband through NSST.Then,the adaptive sparse representation(ASR) model is used to merge the low-frequency sparse coefficients in the NSST domain.At the same time,a parameter-adaptive pulse-coupled neural network(PA-PCNN) model is used to fuse the corresponding high-frequency parts.Finally,the fused image is reconstructed by performing inverse NSST on the fused low-frequency and high-frequency bands.Experimental results demonstrate that the algorithm solves the "block effect" problem of the conventional SR model,overcomes the difficulty of setting free parameters in the conventional PCNN model,leading to state-of-the-art results on both subjective vision and objective assessment.
作者
王昭
杜庆治
董安勇
苏斌
赵文博
WANG Zhao;DU Qing-zhi;DONG An-yong;SU Bin;ZHAO Wen-bo(Kunming University of Science and Technology,Faculty of information engineering and automation,Kunming,650500China;Kunming North Infrared Kunming,650500China)
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2020年第10期1062-1073,共12页
Journal of Optoelectronics·Laser
基金
昆明市科技局科技成果推广应用及科技惠民计划项目《红外热成像森林防火及资源管理系统推广应用》(昆科计字2016-2-G-05372)资助项目。
