基于复合域多尺度分解的红外偏振图像融合方法  

Infrared Polarization Image Fusion Method Based onComposite Domain Multi-scale Decomposition

作　　者：陈广秋[1] 魏洲 段锦[1] 黄丹丹 CHEN Guangqiu;WEI Zhou;DUAN Jin;HUANG Dandan(College of Electronic Information Engineering,Changchun University of Science and Technology,Changchun 130022,China)

机构地区：[1]长春理工大学电子信息工程学院,长春130022

出　　处：《吉林大学学报(理学版)》2025年第2期479-491,共13页Journal of Jilin University:Science Edition

基　　金：国家自然科学基金重大仪器专项基金(批准号:62127813);吉林省科技发展计划项目(批准号:20210203181SF)。

摘　　要：针对目前红外偏振融合图像质量差、偏振信息缺失、目标纹理细节不够等问题,提出一种基于复合域多尺度分解的红外偏振图像融合方法.首先,在空间域内利用引导滤波器对源图像进行二尺度分解,得到细节层和基础层,在频域内利用非下采样剪切波变换对基础层图像进行多尺度多方向分解,得到低频子带图像和高频子带图像;其次,对高频子带采用主成分分析-自适应脉冲耦合神经网络融合规则,对低频子带采用改进的卷积稀疏表示进行系数合并,细节层融合采用基于像素相似度的局部能量加权和选择性融合规则;最后,在复合域内利用逆变换重构出融合图像.实验结果表明,该方法在主观视觉性能和8个客观评价指标上均优于其他对比融合方法,说明该方法在红外偏振图像融合中具有较多优势,能有效提高融合图像的质量.Aiming at the problems of poor image quality,lack of polarization information,and inadequate target texture details in current infrared polarization image fusion,we proposed an infrared polarization image fusion method based on composite domain multi-scale decomposition.Firstly,in the spatial domain,a two-scale decomposition of the source image was performed by using a bootstrap filter to obtain the detail and base layers,in the frequency domain,a multi-scale multi-directional decomposition of the base layer image was performed by using a non-subsampled shear-wave transform to obtain the low-frequency sub-band image and high-frequency sub-band image.Secondly,the principal component analysis-adaptive pulse coupled neural network fusion rule was used for high-frequency sub-band,an improved convolutional sparse representation was used for coefficient merging for the low-frequency sub-bands,and local energy weighting and selective fusion rules based on pixel similarity were used for detail layed fusion.Finally,the fused ima ge was reconstructed by using an inverse transformation in the composite domain.Experimental results show that the proposed method outperforms other comparative fusion methods in subjective visual performance and eight objective evaluation metrics,indicating that the method has many advantages in infrared polarization image fusion and can effectively enhance the quality of fused images.

关 键 词：红外偏振图像融合 非下采样剪切波变换 自适应脉冲耦合神经网络 卷积稀疏表示 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]