DSFA: cross-scene domain style and feature adaptation for landslide detection from high spatial resolution images
Rapid and accurate landslide inventory mapping is significant for emergency rescue and post-disaster reconstruction. Nowadays, deep learning methods exhibit excellent performance in supervised landslide detection. However, due to differences between cross-scene images, the performance of existing me...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Taylor & Francis Group
2023-12-01
|
Series: | International Journal of Digital Earth |
Subjects: | |
Online Access: | http://dx.doi.org/10.1080/17538947.2023.2229794 |
_version_ | 1797678414438596608 |
---|---|
author | Penglei Li Yi Wang Tongzhen Si Kashif Ullah Wei Han Lizhe Wang |
author_facet | Penglei Li Yi Wang Tongzhen Si Kashif Ullah Wei Han Lizhe Wang |
author_sort | Penglei Li |
collection | DOAJ |
description | Rapid and accurate landslide inventory mapping is significant for emergency rescue and post-disaster reconstruction. Nowadays, deep learning methods exhibit excellent performance in supervised landslide detection. However, due to differences between cross-scene images, the performance of existing methods is significantly degraded when directly applied to another scene, which limits the application of rapid landslide inventory mapping. In this study, we propose a novel Domain Style and Feature Adaptation (DSFA) method for cross-scene landslide detection from high spatial resolution images, which can leverage labeled source domain images and unlabeled target domain images to mine robust landslide representations for different scenes. Specifically, we mitigate the large discrepancy between domains at the dataset level and feature level. At the dataset level, we introduce a domain style adaptation strategy to shift landslide styles, which not only bridges the domain gap, but also increases the diversity of landslide samples. At the feature level, adversarial learning and domain distance minimization are integrated to narrow large feature distribution discrepancies for learning domain-invariant information. In addition, to avoid information omission, we improve the U-Net3+ model. Extensive experimental results demonstrate that DSFA has superior detection capability and outperforms other methods, showing its great application potential in unsupervised landslide domain detection. |
first_indexed | 2024-03-11T22:59:26Z |
format | Article |
id | doaj.art-3acd591d131045208eb50a7f6fa1e01c |
institution | Directory Open Access Journal |
issn | 1753-8947 1753-8955 |
language | English |
last_indexed | 2024-03-11T22:59:26Z |
publishDate | 2023-12-01 |
publisher | Taylor & Francis Group |
record_format | Article |
series | International Journal of Digital Earth |
spelling | doaj.art-3acd591d131045208eb50a7f6fa1e01c2023-09-21T15:09:03ZengTaylor & Francis GroupInternational Journal of Digital Earth1753-89471753-89552023-12-011612426244710.1080/17538947.2023.22297942229794DSFA: cross-scene domain style and feature adaptation for landslide detection from high spatial resolution imagesPenglei Li0Yi Wang1Tongzhen Si2Kashif Ullah3Wei Han4Lizhe Wang5School of Geophysics and Geomatics, China University of GeosciencesSchool of Geophysics and Geomatics, China University of GeosciencesSchool of Computer Science, Wuhan UniversitySchool of Geophysics and Geomatics, China University of GeosciencesSchool of Computer Science, China University of GeosciencesSchool of Computer Science, China University of GeosciencesRapid and accurate landslide inventory mapping is significant for emergency rescue and post-disaster reconstruction. Nowadays, deep learning methods exhibit excellent performance in supervised landslide detection. However, due to differences between cross-scene images, the performance of existing methods is significantly degraded when directly applied to another scene, which limits the application of rapid landslide inventory mapping. In this study, we propose a novel Domain Style and Feature Adaptation (DSFA) method for cross-scene landslide detection from high spatial resolution images, which can leverage labeled source domain images and unlabeled target domain images to mine robust landslide representations for different scenes. Specifically, we mitigate the large discrepancy between domains at the dataset level and feature level. At the dataset level, we introduce a domain style adaptation strategy to shift landslide styles, which not only bridges the domain gap, but also increases the diversity of landslide samples. At the feature level, adversarial learning and domain distance minimization are integrated to narrow large feature distribution discrepancies for learning domain-invariant information. In addition, to avoid information omission, we improve the U-Net3+ model. Extensive experimental results demonstrate that DSFA has superior detection capability and outperforms other methods, showing its great application potential in unsupervised landslide domain detection.http://dx.doi.org/10.1080/17538947.2023.2229794landslide detectiondeep learningremote sensingdomain adaptationhigh spatial resolution |
spellingShingle | Penglei Li Yi Wang Tongzhen Si Kashif Ullah Wei Han Lizhe Wang DSFA: cross-scene domain style and feature adaptation for landslide detection from high spatial resolution images International Journal of Digital Earth landslide detection deep learning remote sensing domain adaptation high spatial resolution |
title | DSFA: cross-scene domain style and feature adaptation for landslide detection from high spatial resolution images |
title_full | DSFA: cross-scene domain style and feature adaptation for landslide detection from high spatial resolution images |
title_fullStr | DSFA: cross-scene domain style and feature adaptation for landslide detection from high spatial resolution images |
title_full_unstemmed | DSFA: cross-scene domain style and feature adaptation for landslide detection from high spatial resolution images |
title_short | DSFA: cross-scene domain style and feature adaptation for landslide detection from high spatial resolution images |
title_sort | dsfa cross scene domain style and feature adaptation for landslide detection from high spatial resolution images |
topic | landslide detection deep learning remote sensing domain adaptation high spatial resolution |
url | http://dx.doi.org/10.1080/17538947.2023.2229794 |
work_keys_str_mv | AT pengleili dsfacrossscenedomainstyleandfeatureadaptationforlandslidedetectionfromhighspatialresolutionimages AT yiwang dsfacrossscenedomainstyleandfeatureadaptationforlandslidedetectionfromhighspatialresolutionimages AT tongzhensi dsfacrossscenedomainstyleandfeatureadaptationforlandslidedetectionfromhighspatialresolutionimages AT kashifullah dsfacrossscenedomainstyleandfeatureadaptationforlandslidedetectionfromhighspatialresolutionimages AT weihan dsfacrossscenedomainstyleandfeatureadaptationforlandslidedetectionfromhighspatialresolutionimages AT lizhewang dsfacrossscenedomainstyleandfeatureadaptationforlandslidedetectionfromhighspatialresolutionimages |