Landsat Super-Resolution Enhancement Using Convolution Neural Networks and Sentinel-2 for Training
Landsat is a fundamental data source for understanding historical change and its effect on environmental processes. In this research we test shallow and deep convolution neural networks (CNNs) for Landsat image super-resolution enhancement, trained using Sentinel-2, in three study sites representing...
Published in: | Remote Sensing |
---|---|
Main Authors: | , , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
MDPI AG
2018
|
Subjects: | |
Online Access: | https://doi.org/10.3390/rs10030394 https://doaj.org/article/ad7716d13fe1480e89c336335f3cc9f7 |
id |
ftdoajarticles:oai:doaj.org/article:ad7716d13fe1480e89c336335f3cc9f7 |
---|---|
record_format |
openpolar |
spelling |
ftdoajarticles:oai:doaj.org/article:ad7716d13fe1480e89c336335f3cc9f7 2023-05-15T18:40:28+02:00 Landsat Super-Resolution Enhancement Using Convolution Neural Networks and Sentinel-2 for Training Darren Pouliot Rasim Latifovic Jon Pasher Jason Duffe 2018-03-01T00:00:00Z https://doi.org/10.3390/rs10030394 https://doaj.org/article/ad7716d13fe1480e89c336335f3cc9f7 EN eng MDPI AG http://www.mdpi.com/2072-4292/10/3/394 https://doaj.org/toc/2072-4292 2072-4292 doi:10.3390/rs10030394 https://doaj.org/article/ad7716d13fe1480e89c336335f3cc9f7 Remote Sensing, Vol 10, Iss 3, p 394 (2018) super resolution convolution neural network Landsat Sentinel-2 Science Q article 2018 ftdoajarticles https://doi.org/10.3390/rs10030394 2022-12-31T16:10:43Z Landsat is a fundamental data source for understanding historical change and its effect on environmental processes. In this research we test shallow and deep convolution neural networks (CNNs) for Landsat image super-resolution enhancement, trained using Sentinel-2, in three study sites representing boreal forest, tundra, and cropland/woodland environments. The analysis sought to assess baseline performance and determine the capacity for spatial and temporal extension of the trained CNNs. This is not a data fusion approach and a high-resolution image is only needed to train the CNN. Results show improvement with the deeper network generally achieving better results. For spatial and temporal extension, the deep CNN performed the same or better than the shallow CNN, but at greater computational cost. Results for temporal extension were influenced by change potentiality reducing the performance difference between the shallow and deep CNN. Visual examination revealed sharper images regarding land cover boundaries, linear features, and within-cover textures. The results suggest that spatial enhancement of the Landsat archive is feasible, with optimal performance where CNNs can be trained and applied within the same spatial domain. Future research will assess the enhancement on time series and associated land cover applications. Article in Journal/Newspaper Tundra Directory of Open Access Journals: DOAJ Articles Remote Sensing 10 3 394 |
institution |
Open Polar |
collection |
Directory of Open Access Journals: DOAJ Articles |
op_collection_id |
ftdoajarticles |
language |
English |
topic |
super resolution convolution neural network Landsat Sentinel-2 Science Q |
spellingShingle |
super resolution convolution neural network Landsat Sentinel-2 Science Q Darren Pouliot Rasim Latifovic Jon Pasher Jason Duffe Landsat Super-Resolution Enhancement Using Convolution Neural Networks and Sentinel-2 for Training |
topic_facet |
super resolution convolution neural network Landsat Sentinel-2 Science Q |
description |
Landsat is a fundamental data source for understanding historical change and its effect on environmental processes. In this research we test shallow and deep convolution neural networks (CNNs) for Landsat image super-resolution enhancement, trained using Sentinel-2, in three study sites representing boreal forest, tundra, and cropland/woodland environments. The analysis sought to assess baseline performance and determine the capacity for spatial and temporal extension of the trained CNNs. This is not a data fusion approach and a high-resolution image is only needed to train the CNN. Results show improvement with the deeper network generally achieving better results. For spatial and temporal extension, the deep CNN performed the same or better than the shallow CNN, but at greater computational cost. Results for temporal extension were influenced by change potentiality reducing the performance difference between the shallow and deep CNN. Visual examination revealed sharper images regarding land cover boundaries, linear features, and within-cover textures. The results suggest that spatial enhancement of the Landsat archive is feasible, with optimal performance where CNNs can be trained and applied within the same spatial domain. Future research will assess the enhancement on time series and associated land cover applications. |
format |
Article in Journal/Newspaper |
author |
Darren Pouliot Rasim Latifovic Jon Pasher Jason Duffe |
author_facet |
Darren Pouliot Rasim Latifovic Jon Pasher Jason Duffe |
author_sort |
Darren Pouliot |
title |
Landsat Super-Resolution Enhancement Using Convolution Neural Networks and Sentinel-2 for Training |
title_short |
Landsat Super-Resolution Enhancement Using Convolution Neural Networks and Sentinel-2 for Training |
title_full |
Landsat Super-Resolution Enhancement Using Convolution Neural Networks and Sentinel-2 for Training |
title_fullStr |
Landsat Super-Resolution Enhancement Using Convolution Neural Networks and Sentinel-2 for Training |
title_full_unstemmed |
Landsat Super-Resolution Enhancement Using Convolution Neural Networks and Sentinel-2 for Training |
title_sort |
landsat super-resolution enhancement using convolution neural networks and sentinel-2 for training |
publisher |
MDPI AG |
publishDate |
2018 |
url |
https://doi.org/10.3390/rs10030394 https://doaj.org/article/ad7716d13fe1480e89c336335f3cc9f7 |
genre |
Tundra |
genre_facet |
Tundra |
op_source |
Remote Sensing, Vol 10, Iss 3, p 394 (2018) |
op_relation |
http://www.mdpi.com/2072-4292/10/3/394 https://doaj.org/toc/2072-4292 2072-4292 doi:10.3390/rs10030394 https://doaj.org/article/ad7716d13fe1480e89c336335f3cc9f7 |
op_doi |
https://doi.org/10.3390/rs10030394 |
container_title |
Remote Sensing |
container_volume |
10 |
container_issue |
3 |
container_start_page |
394 |
_version_ |
1766229832035729408 |