基于LatLRR与NSP分解的红外与可见光图像融合北大核心CSCD

针对现有红外与可见光图像融合过程中存在的图像对比度低、红外特征不明显等问题,提出了一种基于非采样金字塔滤波(Nonsubsampled Pyramid,NSP)与潜在低秩表示(Latent Low Rank Representation,LatLRR)分解的红外与可见光图像融合算法。首先,对红外与可见光图像进行分解,采用NSP分解提取源图像的低频信息,LatLRR分解提取源图像的局部结构信息;其次,根据红外低频信息与可见光低频信息的特征及融合结果图像中低频分量占比,利用红外像素强度权重调控策略完成对低频信息的融合,同时,为使红外与可见光的局部结构信息在融合时保持均衡,使用基于像素灰度值求和的策略进行1∶1融合;最后,图像重构中引入非线性变换思想,使局部结构信息与低频信息有更加完美的契合。实验结果表明,融合结果图像在极大保留红外特征的同时又能兼顾可见光图像中的细节信息,该算法能够对红外与可见光图像进行有效融合。

Infrared and visible image fusion basedon LatLRR and NSP decomposition

An infrared and visible image fusion algorithm based on Nonsubsampled Pyramid (NSP) and Latent Low Rank Representation (LatLRR) is proposed to address the problems of low image contrast and inconspicuous infrared features in the existing infrared and visible image fusion process.Firstly,the infrared and visible images are decomposed.The low frequency information of the source image is extracted by NSP decomposition,and the local structure information of the source image is extracted by LatLRR decomposition.Secondly,according to the characteristics of infrared low frequency information and visible low frequency information and the proportion of low frequency components in the fusion result image,the infrared pixel strength weight adjustment strategy is used to complete the fusion of low frequency information.At the same time,in order to maintain the balance of the local structure information of infrared and visible light during the fusion,the 1:1 fusion strategy based on the sum of pixel gray values is used.Finally,the idea of nonlinear transformation is introduced into image reconstruction to make the local structure information and low frequency information more perfectly fit.The experimental results show that the fusion result image can take into account the detail information in the visible image while greatly preserving the infrared features,and the algorithm can effectively fuse the infrared and visible images.