红外技术2026，Vol.48Issue(2)：166-175,10.
基于张量环多模低秩与图正则的遥感图像融合方法
Remote Sensing Image Fusion Method Based on Tensor Ring Multimode Low-Rank Graph Regularization
摘要
Abstract
Remote sensing image fusion is an economical and effective approach for obtaining high-spatial-resolution hyperspectral images,capable of overcoming the limitations of single sensors.However,it involves an ill-posed inverse problem and is susceptible to noise contamination.To address these challenges,this paper proposes an image fusion model based on tensor ring decomposition,transforming the fusion process into the estimation of target image tensor ring factors.Low-dimensional subspace features are further utilized to achieve super-resolution reconstruction of high-dimensional data.First,the local similarity features of the tensor ring factors are exploited by constructing a multimode graph regularization term.Second,an approximation of global low-rank features in low-dimensional subspaces is obtained by introducing tensor nuclear norms for the truncated singular value decomposition of tensor ring factors.Finally,an efficient algorithm was designed to realize model optimization and solution.Experimental results on multiple datasets demonstrate that the proposed fusion model effectively enhances the spatial resolution of hyperspectral images while significantly suppressing noise.关键词
高光谱图像/张量环/遥感图像融合/张量分解/凸优化Key words
hyperspectral image/tensor ring/remote sensing image fusion/tensor decomposition/convex optimization分类
信息技术与安全科学引用本文复制引用
马飞,曲强,杨飞霞,徐光宪..基于张量环多模低秩与图正则的遥感图像融合方法[J].红外技术,2026,48(2):166-175,10.基金项目
辽宁省教育厅高校科研业务经费项目(LJ242410147006) （LJ242410147006）
辽宁工程技术大学GPU资源支持项目. （）
