Blind Fusion of Hyperspectral Multispectral Images Based on Matrix Factorization

The fusion of low spatial resolution hyperspectral images and high spatial resolution multispectral images in the same scenario is important for the super-resolution of hyperspectral images. The spectral response function (SRF) and the point spread function (PSF) are two crucial prior pieces of info...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Jian Long, Yuanxi Peng
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	hyperspectral imaging super-resolution image fusion matrix factorization Science Q
Acceso en línea:	https://doaj.org/article/ef0209aa8eb34278afba97941a12941a
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:ef0209aa8eb34278afba97941a12941a
record_format	dspace
spelling	oai:doaj.org-article:ef0209aa8eb34278afba97941a12941a2021-11-11T18:50:26ZBlind Fusion of Hyperspectral Multispectral Images Based on Matrix Factorization10.3390/rs132142192072-4292https://doaj.org/article/ef0209aa8eb34278afba97941a12941a2021-10-01T00:00:00Zhttps://www.mdpi.com/2072-4292/13/21/4219https://doaj.org/toc/2072-4292The fusion of low spatial resolution hyperspectral images and high spatial resolution multispectral images in the same scenario is important for the super-resolution of hyperspectral images. The spectral response function (SRF) and the point spread function (PSF) are two crucial prior pieces of information in fusion, and most of the current algorithms need to provide these two preliminary pieces of information in advance, even for semi-blind fusion algorithms at least the SRF. This causes limitations in the application of fusion algorithms. This paper aims to solve the dependence of the fusion method on the point spread function and proposes a method to estimate the spectral response function from the images involved in the fusion to achieve blind fusion. We conducted experiments on simulated datasets Pavia University, CAVE, and the remote sensing images acquired by two spectral cameras, Sentinel 2 and Hyperion. The experimental results show that our proposed SRF estimation method can improve the PSNR value by 5 dB on average compared with other state-of-the-art SRF estimation results. The proposed blind fusion method can improve the PSNR value of fusion results by 3–15 dB compared with other blind fusion methods.Jian LongYuanxi PengMDPI AGarticlehyperspectral imaging super-resolutionimage fusionmatrix factorizationScienceQENRemote Sensing, Vol 13, Iss 4219, p 4219 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	hyperspectral imaging super-resolution image fusion matrix factorization Science Q
spellingShingle	hyperspectral imaging super-resolution image fusion matrix factorization Science Q Jian Long Yuanxi Peng Blind Fusion of Hyperspectral Multispectral Images Based on Matrix Factorization
description	The fusion of low spatial resolution hyperspectral images and high spatial resolution multispectral images in the same scenario is important for the super-resolution of hyperspectral images. The spectral response function (SRF) and the point spread function (PSF) are two crucial prior pieces of information in fusion, and most of the current algorithms need to provide these two preliminary pieces of information in advance, even for semi-blind fusion algorithms at least the SRF. This causes limitations in the application of fusion algorithms. This paper aims to solve the dependence of the fusion method on the point spread function and proposes a method to estimate the spectral response function from the images involved in the fusion to achieve blind fusion. We conducted experiments on simulated datasets Pavia University, CAVE, and the remote sensing images acquired by two spectral cameras, Sentinel 2 and Hyperion. The experimental results show that our proposed SRF estimation method can improve the PSNR value by 5 dB on average compared with other state-of-the-art SRF estimation results. The proposed blind fusion method can improve the PSNR value of fusion results by 3–15 dB compared with other blind fusion methods.
format	article
author	Jian Long Yuanxi Peng
author_facet	Jian Long Yuanxi Peng
author_sort	Jian Long
title	Blind Fusion of Hyperspectral Multispectral Images Based on Matrix Factorization
title_short	Blind Fusion of Hyperspectral Multispectral Images Based on Matrix Factorization
title_full	Blind Fusion of Hyperspectral Multispectral Images Based on Matrix Factorization
title_fullStr	Blind Fusion of Hyperspectral Multispectral Images Based on Matrix Factorization
title_full_unstemmed	Blind Fusion of Hyperspectral Multispectral Images Based on Matrix Factorization
title_sort	blind fusion of hyperspectral multispectral images based on matrix factorization
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/ef0209aa8eb34278afba97941a12941a
work_keys_str_mv	AT jianlong blindfusionofhyperspectralmultispectralimagesbasedonmatrixfactorization AT yuanxipeng blindfusionofhyperspectralmultispectralimagesbasedonmatrixfactorization
_version_	1718431738979418112

Blind Fusion of Hyperspectral Multispectral Images Based on Matrix Factorization

Ejemplares similares