Rapid Mapping of Landslides on SAR Data by Attention U-Net

Multiple landslide events are common around the globe. They can cause severe damage to both human lives and infrastructures. Although a huge quantity of research has been shaped to address rapid mapping of landslides by optical Earth Observation (EO) data, various gaps and uncertainties are still pr...

Descripción completa

Detalles Bibliográficos
Autores: Nava, L, Bhuyan, K, Meena, SR, Monserrat, O, Catani, F
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2022
País:España
Institución:Centre Tecnològic de Telecomunicacions de Catalunya (CTTC)
Repositorio:r-CTTC. Repositorio Institucional Producción Científica del Centre Tecnològic de Telecomunicacions de Catalunya (CTTC)
OAI Identifier:oai:cttc.fundanetsuite.com:p6430
Acceso en línea:https://cttc.fundanetsuite.com/Publicaciones/ProdCientif/PublicacionFrw.aspx?id=6430
Access Level:acceso abierto
Palabra clave:landslides
SAR
sentinel-1
deep learning
convolutional neural network
U-Net
id ES_4e4a0ade9b5e731efe06795c8fe31daa
oai_identifier_str oai:cttc.fundanetsuite.com:p6430
network_acronym_str ES
network_name_str España
repository_id_str
spelling Rapid Mapping of Landslides on SAR Data by Attention U-NetNava, LBhuyan, KMeena, SRMonserrat, OCatani, FlandslidesSARsentinel-1deep learningconvolutional neural networkU-NetMultiple landslide events are common around the globe. They can cause severe damage to both human lives and infrastructures. Although a huge quantity of research has been shaped to address rapid mapping of landslides by optical Earth Observation (EO) data, various gaps and uncertainties are still present when dealing with cloud obscuration and 24/7 operativity. To address the issue, we explore the usage of SAR data over the eastern Iburi sub-prefecture of Hokkaido, Japan. In the area, about 8000 co-seismic landslides were triggered by an Mw 6.6 earthquake on 6 September 2018, at 03.08 local time (JST). In the following study, we modify a Deep Learning (DL) convolutional neural network (CNN) architecture suited for pixel-based classification purposes, the so-called Attention U-Net (Attn-U-Net) and we employ it to evaluate the potential of bi- and tri-temporal SAR amplitude data from the Sentinel-1 satellite and slope angle to map landslides even under thick cloud cover. Four different datasets, composed of two different band combinations per two satellite orbits (ascending and descending) are analyzed. Moreover, the impact of augmentations is evaluated independently for each dataset. The models' predictions are compared against an accurate landslide inventory obtained by manual mapping on pre-and post-event PlanetScope imagery through F1-score and other common metrics. The best result was yielded by the augmented ascending tri-temporal SAR composite image (61% F1-score). Augmentations have a positive impact on the ascending Sentinel-1 orbit, while metrics decrease when augmentations are applied on descending path. Our findings demonstrate that combining SAR data with other data sources may help to map landslides quickly, even during storms and under deep cloud cover. However, further investigations and improvements are still needed, this being one of the first attempts in which the combination of SAR data and DL algorithms are employed for landslide mapping purposes.MDPI Multidisciplinary Digital Publishing Institute2022info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersionhttps://cttc.fundanetsuite.com/Publicaciones/ProdCientif/PublicacionFrw.aspx?id=6430Remote SensingISSN: 20724292reponame:r-CTTC. Repositorio Institucional Producción Científica del Centre Tecnològic de Telecomunicacions de Catalunya (CTTC)instname:Centre Tecnològic de Telecomunicacions de Catalunya (CTTC)Inglésinfo:eu-repo/semantics/openAccessoai:cttc.fundanetsuite.com:p64302026-06-17T11:44:47Z
dc.title.none.fl_str_mv Rapid Mapping of Landslides on SAR Data by Attention U-Net
title Rapid Mapping of Landslides on SAR Data by Attention U-Net
spellingShingle Rapid Mapping of Landslides on SAR Data by Attention U-Net
Nava, L
landslides
SAR
sentinel-1
deep learning
convolutional neural network
U-Net
title_short Rapid Mapping of Landslides on SAR Data by Attention U-Net
title_full Rapid Mapping of Landslides on SAR Data by Attention U-Net
title_fullStr Rapid Mapping of Landslides on SAR Data by Attention U-Net
title_full_unstemmed Rapid Mapping of Landslides on SAR Data by Attention U-Net
title_sort Rapid Mapping of Landslides on SAR Data by Attention U-Net
dc.creator.none.fl_str_mv Nava, L
Bhuyan, K
Meena, SR
Monserrat, O
Catani, F
author Nava, L
author_facet Nava, L
Bhuyan, K
Meena, SR
Monserrat, O
Catani, F
author_role author
author2 Bhuyan, K
Meena, SR
Monserrat, O
Catani, F
author2_role author
author
author
author
dc.subject.none.fl_str_mv landslides
SAR
sentinel-1
deep learning
convolutional neural network
U-Net
topic landslides
SAR
sentinel-1
deep learning
convolutional neural network
U-Net
description Multiple landslide events are common around the globe. They can cause severe damage to both human lives and infrastructures. Although a huge quantity of research has been shaped to address rapid mapping of landslides by optical Earth Observation (EO) data, various gaps and uncertainties are still present when dealing with cloud obscuration and 24/7 operativity. To address the issue, we explore the usage of SAR data over the eastern Iburi sub-prefecture of Hokkaido, Japan. In the area, about 8000 co-seismic landslides were triggered by an Mw 6.6 earthquake on 6 September 2018, at 03.08 local time (JST). In the following study, we modify a Deep Learning (DL) convolutional neural network (CNN) architecture suited for pixel-based classification purposes, the so-called Attention U-Net (Attn-U-Net) and we employ it to evaluate the potential of bi- and tri-temporal SAR amplitude data from the Sentinel-1 satellite and slope angle to map landslides even under thick cloud cover. Four different datasets, composed of two different band combinations per two satellite orbits (ascending and descending) are analyzed. Moreover, the impact of augmentations is evaluated independently for each dataset. The models' predictions are compared against an accurate landslide inventory obtained by manual mapping on pre-and post-event PlanetScope imagery through F1-score and other common metrics. The best result was yielded by the augmented ascending tri-temporal SAR composite image (61% F1-score). Augmentations have a positive impact on the ascending Sentinel-1 orbit, while metrics decrease when augmentations are applied on descending path. Our findings demonstrate that combining SAR data with other data sources may help to map landslides quickly, even during storms and under deep cloud cover. However, further investigations and improvements are still needed, this being one of the first attempts in which the combination of SAR data and DL algorithms are employed for landslide mapping purposes.
publishDate 2022
dc.date.none.fl_str_mv 2022
dc.type.none.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
format article
status_str publishedVersion
dc.identifier.none.fl_str_mv https://cttc.fundanetsuite.com/Publicaciones/ProdCientif/PublicacionFrw.aspx?id=6430
url https://cttc.fundanetsuite.com/Publicaciones/ProdCientif/PublicacionFrw.aspx?id=6430
dc.language.none.fl_str_mv Inglés
language_invalid_str_mv Inglés
dc.rights.none.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.publisher.none.fl_str_mv MDPI Multidisciplinary Digital Publishing Institute
publisher.none.fl_str_mv MDPI Multidisciplinary Digital Publishing Institute
dc.source.none.fl_str_mv Remote Sensing
ISSN: 20724292
reponame:r-CTTC. Repositorio Institucional Producción Científica del Centre Tecnològic de Telecomunicacions de Catalunya (CTTC)
instname:Centre Tecnològic de Telecomunicacions de Catalunya (CTTC)
instname_str Centre Tecnològic de Telecomunicacions de Catalunya (CTTC)
reponame_str r-CTTC. Repositorio Institucional Producción Científica del Centre Tecnològic de Telecomunicacions de Catalunya (CTTC)
collection r-CTTC. Repositorio Institucional Producción Científica del Centre Tecnològic de Telecomunicacions de Catalunya (CTTC)
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869407746637430784
score 15,811543

Similares en LA Referencia