Publication

Using LUCAS survey and Recurrent Neural Networks to produce LCLU classification based on a Satellite Image time series of Sentinel-2

View document

Bibliographic Details
Summary:The need of timely and accurate information for the territory has increased over the years, making Land Cover Land Use (LCLU) mapping one of the most common application of remote sensing. Recently, the advances in satellite technology and the open access policies for remote sensing data increased the interest in exploring satellite image time series. In addition, the attention of researchers has shifted from standard machine learning algorithms (e.g., Support Vector Machines and Random Forest) to Recurrent Neural Networks due to their ability of exploiting sequential information. However, acquiring reference data to train these algorithms is still a hurdle. This study aims to evaluate the capability of a Gated Recurrent Unit in performing pixel-level LCLU classification of a satellite image time series, using Sentinel-2 imagery and having the LUCAS survey as reference data. To assess the performance of our model we compared it to state-of-the-art classifiers (SVM and RF). Due to the unbalance nature of the LUCAS survey, we applied oversampling to this dataset to increase the performance of our models, testing three different oversampling techniques. The results attained showed that Recurrent Neural Networks did not outperform the other state-of-the-art algorithms, when trained with a limited number of sampling units, and that oversampling the LUCAS survey increased the performance of all the classifiers. Finally, we were able to demonstrate that it is possible to produce LCLU classification of satellite image time series using only open-source data by using Sentinel-2 imagery and the LUCAS survey as refence data.
Main Authors:Silva, Nuno Alexandre Pereira da
Subject:LCLU classification LUCAS survey Recurrent Neural Networks Oversampling Sentinel-2
Year:2021
Country:Portugal
Document type:master thesis
Access type:open access
Associated institution:Universidade Nova de Lisboa
Language:English
Origin:Repositório Institucional da UNL
_version_ 1868983616808157184
author Silva, Nuno Alexandre Pereira da
author_facet Silva, Nuno Alexandre Pereira da
author_role author
contributor_name_str_mv Caetano, Mário Sílvio Rochinha de Andrade
Castelli, Mauro
RUN
country_str PT
creators_json_txt [{\"Person.name\":\"Silva, Nuno Alexandre Pereira da\"}]
datacite.contributors.contributor.contributorName.fl_str_mv Caetano, Mário Sílvio Rochinha de Andrade
Castelli, Mauro
RUN
datacite.creators.creator.creatorName.fl_str_mv Silva, Nuno Alexandre Pereira da
datacite.date.Accepted.fl_str_mv 2021-06-22T00:00:00Z
datacite.date.available.fl_str_mv 2021-08-23T16:14:30Z
datacite.date.embargoed.fl_str_mv 2021-08-23T16:14:30Z
datacite.rights.fl_str_mv http://purl.org/coar/access_right/c_abf2
datacite.subjects.subject.fl_str_mv LCLU classification
LUCAS survey
Recurrent Neural Networks
Oversampling
Sentinel-2
datacite.titles.title.fl_str_mv Using LUCAS survey and Recurrent Neural Networks to produce LCLU classification based on a Satellite Image time series of Sentinel-2
dc.contributor.none.fl_str_mv Caetano, Mário Sílvio Rochinha de Andrade
Castelli, Mauro
RUN
dc.creator.none.fl_str_mv Silva, Nuno Alexandre Pereira da
dc.date.Accepted.fl_str_mv 2021-06-22T00:00:00Z
dc.date.available.fl_str_mv 2021-08-23T16:14:30Z
dc.date.embargoed.fl_str_mv 2021-08-23T16:14:30Z
dc.format.none.fl_str_mv application/pdf
dc.identifier.none.fl_str_mv http://hdl.handle.net/10362/122980
dc.language.none.fl_str_mv eng
dc.rights.cclincense.fl_str_mv http://creativecommons.org/licenses/by/4.0/
dc.rights.none.fl_str_mv http://purl.org/coar/access_right/c_abf2
dc.subject.none.fl_str_mv LCLU classification
LUCAS survey
Recurrent Neural Networks
Oversampling
Sentinel-2
dc.title.fl_str_mv Using LUCAS survey and Recurrent Neural Networks to produce LCLU classification based on a Satellite Image time series of Sentinel-2
dc.type.none.fl_str_mv http://purl.org/coar/resource_type/c_bdcc
description The need of timely and accurate information for the territory has increased over the years, making Land Cover Land Use (LCLU) mapping one of the most common application of remote sensing. Recently, the advances in satellite technology and the open access policies for remote sensing data increased the interest in exploring satellite image time series. In addition, the attention of researchers has shifted from standard machine learning algorithms (e.g., Support Vector Machines and Random Forest) to Recurrent Neural Networks due to their ability of exploiting sequential information. However, acquiring reference data to train these algorithms is still a hurdle. This study aims to evaluate the capability of a Gated Recurrent Unit in performing pixel-level LCLU classification of a satellite image time series, using Sentinel-2 imagery and having the LUCAS survey as reference data. To assess the performance of our model we compared it to state-of-the-art classifiers (SVM and RF). Due to the unbalance nature of the LUCAS survey, we applied oversampling to this dataset to increase the performance of our models, testing three different oversampling techniques. The results attained showed that Recurrent Neural Networks did not outperform the other state-of-the-art algorithms, when trained with a limited number of sampling units, and that oversampling the LUCAS survey increased the performance of all the classifiers. Finally, we were able to demonstrate that it is possible to produce LCLU classification of satellite image time series using only open-source data by using Sentinel-2 imagery and the LUCAS survey as refence data.
dirty 0
eu_rights_str_mv openAccess
format masterThesis
fulltext.url.fl_str_mv https://run.unl.pt/bitstreams/c3791026-1d47-4f9a-92b4-0231d4bf9800/download
id run_8aee2d0a8ce8e79137c83e47396ba2bc
identifier.url.fl_str_mv http://hdl.handle.net/10362/122980
inst_facet_str urn:organizationAcronym:unl{{{_:::_}}}Universidade Nova de Lisboa
instacron_str unl
institution Universidade Nova de Lisboa
instname_str Universidade Nova de Lisboa
language eng
network_acronym_str run
network_name_str Repositório Institucional da UNL
oai_identifier_str oai:run.unl.pt:10362/122980
organization_str_mv urn:organizationAcronym:unl
person_str_mv Silva, Nuno Alexandre Pereira da
publishDate 2021
repo_facet_str urn:repositoryAcronym:run{{{_:::_}}}Repositório Institucional da UNL
reponame_str Repositório Institucional da UNL
repository_id_str urn:repositoryAcronym:run
service_str_mv urn:repositoryAcronym:run
spelling engpt_PTThe need of timely and accurate information for the territory has increased over the years, making Land Cover Land Use (LCLU) mapping one of the most common application of remote sensing. Recently, the advances in satellite technology and the open access policies for remote sensing data increased the interest in exploring satellite image time series. In addition, the attention of researchers has shifted from standard machine learning algorithms (e.g., Support Vector Machines and Random Forest) to Recurrent Neural Networks due to their ability of exploiting sequential information. However, acquiring reference data to train these algorithms is still a hurdle. This study aims to evaluate the capability of a Gated Recurrent Unit in performing pixel-level LCLU classification of a satellite image time series, using Sentinel-2 imagery and having the LUCAS survey as reference data. To assess the performance of our model we compared it to state-of-the-art classifiers (SVM and RF). Due to the unbalance nature of the LUCAS survey, we applied oversampling to this dataset to increase the performance of our models, testing three different oversampling techniques. The results attained showed that Recurrent Neural Networks did not outperform the other state-of-the-art algorithms, when trained with a limited number of sampling units, and that oversampling the LUCAS survey increased the performance of all the classifiers. Finally, we were able to demonstrate that it is possible to produce LCLU classification of satellite image time series using only open-source data by using Sentinel-2 imagery and the LUCAS survey as refence data.application/pdfpt_PTUsing LUCAS survey and Recurrent Neural Networks to produce LCLU classification based on a Satellite Image time series of Sentinel-2Silva, Nuno Alexandre Pereira daCaetano, Mário Sílvio Rochinha de AndradeCastelli, MauroHostingInstitutionOrganizationalRUNe-mailmailto:run@unl.ptrun@unl.ptURNurn:tid:2027571962021-08-23T16:14:30Z2021-06-222021-06-22T00:00:00ZHandlehttp://hdl.handle.net/10362/122980http://purl.org/coar/access_right/c_abf2open accessLCLU classificationLUCAS surveyRecurrent Neural NetworksOversamplingSentinel-21134671 bytesliteraturehttp://purl.org/coar/resource_type/c_bdccmaster thesis2021-06-22http://creativecommons.org/licenses/by/4.0/http://purl.org/coar/access_right/c_abf2application/pdffulltexthttps://run.unl.pt/bitstreams/c3791026-1d47-4f9a-92b4-0231d4bf9800/download
spellingShingle Using LUCAS survey and Recurrent Neural Networks to produce LCLU classification based on a Satellite Image time series of Sentinel-2
Silva, Nuno Alexandre Pereira da
LCLU classification
LUCAS survey
Recurrent Neural Networks
Oversampling
Sentinel-2
status SINGLETON
subject.fl_str_mv LCLU classification
LUCAS survey
Recurrent Neural Networks
Oversampling
Sentinel-2
title Using LUCAS survey and Recurrent Neural Networks to produce LCLU classification based on a Satellite Image time series of Sentinel-2
title_full Using LUCAS survey and Recurrent Neural Networks to produce LCLU classification based on a Satellite Image time series of Sentinel-2
title_fullStr Using LUCAS survey and Recurrent Neural Networks to produce LCLU classification based on a Satellite Image time series of Sentinel-2
title_full_unstemmed Using LUCAS survey and Recurrent Neural Networks to produce LCLU classification based on a Satellite Image time series of Sentinel-2
title_short Using LUCAS survey and Recurrent Neural Networks to produce LCLU classification based on a Satellite Image time series of Sentinel-2
title_sort Using LUCAS survey and Recurrent Neural Networks to produce LCLU classification based on a Satellite Image time series of Sentinel-2
topic LCLU classification
LUCAS survey
Recurrent Neural Networks
Oversampling
Sentinel-2
topic_facet LCLU classification
LUCAS survey
Recurrent Neural Networks
Oversampling
Sentinel-2
url http://hdl.handle.net/10362/122980
visible 1