Fast Super-Resolution of 20 m Sentinel-2 Bands Using Convolutional Neural Networks

Images provided by the ESA Sentinel-2 mission are rapidly becoming the main source of information for the entire remote sensing community, thanks to their unprecedented combination of spatial, spectral and temporal resolution, as well as their associated open access policy. Due to a sensor design tr...

Full description

Bibliographic Details
Main Authors: Massimiliano Gargiulo, Antonio Mazza, Raffaele Gaetano, Giuseppe Ruello, Giuseppe Scarpa
Format: Article
Language:English
Published: MDPI AG 2019-11-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/11/22/2635
id doaj-30209dcc5f07499fb44ad60199af7a6b
record_format Article
spelling doaj-30209dcc5f07499fb44ad60199af7a6b2020-11-25T00:04:25ZengMDPI AGRemote Sensing2072-42922019-11-011122263510.3390/rs11222635rs11222635Fast Super-Resolution of 20 m Sentinel-2 Bands Using Convolutional Neural NetworksMassimiliano Gargiulo0Antonio Mazza1Raffaele Gaetano2Giuseppe Ruello3Giuseppe Scarpa4Department of Electrical Engineering and Information Technology (DIETI), University Federico II, 80125 Naples, ItalyDepartment of Electrical Engineering and Information Technology (DIETI), University Federico II, 80125 Naples, ItalyCentre International de Recherche Agronomique pour le Développement (CIRAD), Unité Mixte de Recherche Territoires, Environnement, Télédétéction et Information Spatiale (UMR TETIS), Maison de la Télédétéction, 34000 Montpellier, FranceDepartment of Electrical Engineering and Information Technology (DIETI), University Federico II, 80125 Naples, ItalyDepartment of Electrical Engineering and Information Technology (DIETI), University Federico II, 80125 Naples, ItalyImages provided by the ESA Sentinel-2 mission are rapidly becoming the main source of information for the entire remote sensing community, thanks to their unprecedented combination of spatial, spectral and temporal resolution, as well as their associated open access policy. Due to a sensor design trade-off, images are acquired (and delivered) at different spatial resolutions (10, 20 and 60 m) according to specific sets of wavelengths, with only the four visible and near infrared bands provided at the highest resolution (10 m). Although this is not a limiting factor in general, many applications seem to emerge in which the resolution enhancement of 20 m bands may be beneficial, motivating the development of specific super-resolution methods. In this work, we propose to leverage Convolutional Neural Networks (CNNs) to provide a fast, upscalable method for the single-sensor fusion of Sentinel-2 (S2) data, whose aim is to provide a 10 m super-resolution of the original 20 m bands. Experimental results demonstrate that the proposed solution can achieve better performance with respect to most of the state-of-the-art methods, including other deep learning based ones with a considerable saving of computational burden.https://www.mdpi.com/2072-4292/11/22/2635pansharpeningdata fusionconvolutional neural networkmulti-resolution analysislandcover classification
collection DOAJ
language English
format Article
sources DOAJ
author Massimiliano Gargiulo
Antonio Mazza
Raffaele Gaetano
Giuseppe Ruello
Giuseppe Scarpa
spellingShingle Massimiliano Gargiulo
Antonio Mazza
Raffaele Gaetano
Giuseppe Ruello
Giuseppe Scarpa
Fast Super-Resolution of 20 m Sentinel-2 Bands Using Convolutional Neural Networks
Remote Sensing
pansharpening
data fusion
convolutional neural network
multi-resolution analysis
landcover classification
author_facet Massimiliano Gargiulo
Antonio Mazza
Raffaele Gaetano
Giuseppe Ruello
Giuseppe Scarpa
author_sort Massimiliano Gargiulo
title Fast Super-Resolution of 20 m Sentinel-2 Bands Using Convolutional Neural Networks
title_short Fast Super-Resolution of 20 m Sentinel-2 Bands Using Convolutional Neural Networks
title_full Fast Super-Resolution of 20 m Sentinel-2 Bands Using Convolutional Neural Networks
title_fullStr Fast Super-Resolution of 20 m Sentinel-2 Bands Using Convolutional Neural Networks
title_full_unstemmed Fast Super-Resolution of 20 m Sentinel-2 Bands Using Convolutional Neural Networks
title_sort fast super-resolution of 20 m sentinel-2 bands using convolutional neural networks
publisher MDPI AG
series Remote Sensing
issn 2072-4292
publishDate 2019-11-01
description Images provided by the ESA Sentinel-2 mission are rapidly becoming the main source of information for the entire remote sensing community, thanks to their unprecedented combination of spatial, spectral and temporal resolution, as well as their associated open access policy. Due to a sensor design trade-off, images are acquired (and delivered) at different spatial resolutions (10, 20 and 60 m) according to specific sets of wavelengths, with only the four visible and near infrared bands provided at the highest resolution (10 m). Although this is not a limiting factor in general, many applications seem to emerge in which the resolution enhancement of 20 m bands may be beneficial, motivating the development of specific super-resolution methods. In this work, we propose to leverage Convolutional Neural Networks (CNNs) to provide a fast, upscalable method for the single-sensor fusion of Sentinel-2 (S2) data, whose aim is to provide a 10 m super-resolution of the original 20 m bands. Experimental results demonstrate that the proposed solution can achieve better performance with respect to most of the state-of-the-art methods, including other deep learning based ones with a considerable saving of computational burden.
topic pansharpening
data fusion
convolutional neural network
multi-resolution analysis
landcover classification
url https://www.mdpi.com/2072-4292/11/22/2635
work_keys_str_mv AT massimilianogargiulo fastsuperresolutionof20msentinel2bandsusingconvolutionalneuralnetworks
AT antoniomazza fastsuperresolutionof20msentinel2bandsusingconvolutionalneuralnetworks
AT raffaelegaetano fastsuperresolutionof20msentinel2bandsusingconvolutionalneuralnetworks
AT giusepperuello fastsuperresolutionof20msentinel2bandsusingconvolutionalneuralnetworks
AT giuseppescarpa fastsuperresolutionof20msentinel2bandsusingconvolutionalneuralnetworks
_version_ 1725429374536122368