张量组稀疏表示的高光谱图像去噪算法

Hyperspectral Image Denoising Based on Tensor Group Sparse Representation

提出了一种基于张量组稀疏表示的高光谱遥感影像降噪。高光谱影像数据可视为三阶张量。首先,高光谱图像被划分为小的张量分块,然后,对相似的张量分块进行聚类,并对聚类分组进行稀疏表示。基于高光谱图像的空间非局部自相似性和光谱相关性,将张量组稀疏表示模型分解为一系列无约束低秩张量的近似问题,进而通过张量分解进行求解。对模拟和真实高光谱数据进行试验,验证了该算法的有效性。

A novel algorithm for hyperspectral image （HSI） denoising is proposed based on tensor group sparse representation. A HSl is considering as 3 order tensor. First, a HSI is divided into small tensor blocks. Second, similar blocks are gathered into clusters, and then a tensor group sparse representation model is constructed based on every cluster. Through exploiting HSI spectral correlation and nonlocal similarity over space, the model constrained tensor group sparse representation can be decomposed into a series of unconstrained low-rank tensor approximation problems, which can be solved using the tensor decomposition technique. The experiment results on the synthetic and real hyperspectral remote sensing images demonstrate the effectiveness of the proposed approach.