%0 Journal Article
%T 基于稀疏表示和光谱自回归的多光谱和高光谱图像融合<br>Multispectral and Hyperspectral Image Fusion Based on Sparse Representation and Spectral Autoregression
%A 钱金明
%J Modeling and Simulation
%P 3196-3211
%@ 2324-870X
%D 2024
%I Hans Publishing
%R 10.12677/mos.2024.133292
%X 现有的高光谱图像融合算法采用将空间维度与光谱维度进行分开融合重建的方式，而高光谱图像光谱维度存在大量的光谱信息，这些像素信息比空间上的像素更加接近目标像素，对于空间维度的重建非常关键。因此，本文提出一种将光谱像素用于对空间维度进行空间信息修复的融合方法，称作基于稀疏表示和光谱自回归的多光谱和高光谱图像融合。该方法通过将光谱维度上的像素，通过自回归模型将其用于空间维度的信息修复，自回归模型由低分辨率的高光谱图像中学习得到，该模型自然保持了光谱间的关联性，通过在公开数据集上的验证了提出模型的有效性。<br />
Existing hyperspectral image fusion algorithms use separate fusion reconstruction of spatial and spectral dimensions, whereas there is a large amount of spectral information in the spectral dimension of hyperspectral images, and the information of these pixels is much closer to the target pixels than spatially similar pixels, which is very critical for the reconstruction of the spatial dimension. Therefore, in this paper, we propose a method to use spectral pixels for spatial information restoration on the spatial dimension, called multispectral and hyperspectral image fusion based on sparse representation and spectral autoregression. The method works by taking the pixels on the spectral dimension and using them for information restoration on the spatial dimension through an autoregressive model, the autoregressive model is learned from low-resolution hyperspectral images, which naturally maintains the correlation between the spectra, and the validity of the proposed model is verified by the validation on the publicly available dataset.
%K 图像融合，高光谱图像，稀疏表示，光谱自回归<br>Image Fusion
%K Hyperspectral Image
%K Sparse Representation
%K Spectral Autoregression
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=87944