Deep Learning-Based Spatial Prediction of Landslide Risk in Coastal Areas Using GIS and Multicriteria Decision Making: A DeepLabV3+ Approach

Sustainable land-use planning and catastrophe risk reduction depend critically on landslide susceptibility mapping. The complex, nonlinear interconnections of environmental and human elements cause terrain instability and challenge conventional prediction methods. In this work, we offer a DeepLabV3+...

Full description

Saved in:
Bibliographic Details
Main Authors: Huyong Yan, Asad Khan, Ahsan Jamil, Belkendil Abdeldjalil, Taoufik Saidani, Nazih Y. Rebouh
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11030867/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850111055617851392
author Huyong Yan
Asad Khan
Ahsan Jamil
Belkendil Abdeldjalil
Taoufik Saidani
Nazih Y. Rebouh
author_facet Huyong Yan
Asad Khan
Ahsan Jamil
Belkendil Abdeldjalil
Taoufik Saidani
Nazih Y. Rebouh
author_sort Huyong Yan
collection DOAJ
description Sustainable land-use planning and catastrophe risk reduction depend critically on landslide susceptibility mapping. The complex, nonlinear interconnections of environmental and human elements cause terrain instability and challenge conventional prediction methods. In this work, we offer a DeepLabV3+-based deep learning framework coupled with geographic information systems and multicriteria decision making methods for spatial prediction of landslide risk, over the Dubai coastal and urban region (covering approximately 4000 km<sup>2</sup>). The approach uses an annotated dataset for semantic segmentation and high-resolution satellite images from the Mohammed Bin Rashid Space Center. On Google Colab with GPU acceleration, the model is trained and verified and then further improved for computational efficiency on a Mac M1 machine. Our results show an overall accuracy of 91.3&#x0025;, a mean intersection over union of 82.5&#x0025;, and an F1-score of 88.4&#x0025;, demonstrating strong classification performance throughout a range of land cover types. The confusion matrix analysis highlights strong segmentation accuracy for water bodies (94.2&#x0025;) and structures (92.4&#x0025;). Considerable misclassification between roadways and unpaved terrain results from spectral similarities. Furthermore, the perclass Dice Coefficient analysis confirms that the model can efficiently discriminate intricate topographical patterns. Especially in fast-expanding areas such as Dubai, UAE, it provides a scalable solution for landslide susceptibility mapping, catastrophe risk management, and sustainable urban design. Future work will explore multisensor data fusion, real-time inference, and applying explainable artificial intelligence techniques to enhance model interpretability in dynamic terrain settings.
format Article
id doaj-art-a70d5fc685f54cd9b6532ec5504aaa68
institution OA Journals
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-a70d5fc685f54cd9b6532ec5504aaa682025-08-20T02:37:42ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118152221523510.1109/JSTARS.2025.357882211030867Deep Learning-Based Spatial Prediction of Landslide Risk in Coastal Areas Using GIS and Multicriteria Decision Making: A DeepLabV3+ ApproachHuyong Yan0Asad Khan1https://orcid.org/0000-0002-1261-0418Ahsan Jamil2https://orcid.org/0000-0001-6855-4147Belkendil Abdeldjalil3https://orcid.org/0000-0001-9247-848XTaoufik Saidani4Nazih Y. Rebouh5https://orcid.org/0000-0002-8621-6595School of Computer Science and Information Engineering, Chongqing Technology and Business University, Chongqing, ChinaMetaverse Research Institute, School of Computer Science and Cyber Engineering, Guangzhou University, Guangzhou, PR ChinaDepartment of Plant and Environmental Sciences, New Mexico State University, Las Cruces, NM, USACenter for Research in Territory Planning (CRAT), Zouaghi Slimane Campus, Constantine, AlgeriaCenter for Scientific Research and Entrepreneurship, Northern Border University, Arar, Saudi ArabiaDepartment of Environmental Management, Institute of Environmental Engineering, RUDN University, Moscow, RussiaSustainable land-use planning and catastrophe risk reduction depend critically on landslide susceptibility mapping. The complex, nonlinear interconnections of environmental and human elements cause terrain instability and challenge conventional prediction methods. In this work, we offer a DeepLabV3+-based deep learning framework coupled with geographic information systems and multicriteria decision making methods for spatial prediction of landslide risk, over the Dubai coastal and urban region (covering approximately 4000 km<sup>2</sup>). The approach uses an annotated dataset for semantic segmentation and high-resolution satellite images from the Mohammed Bin Rashid Space Center. On Google Colab with GPU acceleration, the model is trained and verified and then further improved for computational efficiency on a Mac M1 machine. Our results show an overall accuracy of 91.3&#x0025;, a mean intersection over union of 82.5&#x0025;, and an F1-score of 88.4&#x0025;, demonstrating strong classification performance throughout a range of land cover types. The confusion matrix analysis highlights strong segmentation accuracy for water bodies (94.2&#x0025;) and structures (92.4&#x0025;). Considerable misclassification between roadways and unpaved terrain results from spectral similarities. Furthermore, the perclass Dice Coefficient analysis confirms that the model can efficiently discriminate intricate topographical patterns. Especially in fast-expanding areas such as Dubai, UAE, it provides a scalable solution for landslide susceptibility mapping, catastrophe risk management, and sustainable urban design. Future work will explore multisensor data fusion, real-time inference, and applying explainable artificial intelligence techniques to enhance model interpretability in dynamic terrain settings.https://ieeexplore.ieee.org/document/11030867/Coastal monitoringDeepLabV3+geographic information systems (GISs)landslide susceptibility mappingmultisensor data fusionremote sensing
spellingShingle Huyong Yan
Asad Khan
Ahsan Jamil
Belkendil Abdeldjalil
Taoufik Saidani
Nazih Y. Rebouh
Deep Learning-Based Spatial Prediction of Landslide Risk in Coastal Areas Using GIS and Multicriteria Decision Making: A DeepLabV3+ Approach
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Coastal monitoring
DeepLabV3+
geographic information systems (GISs)
landslide susceptibility mapping
multisensor data fusion
remote sensing
title Deep Learning-Based Spatial Prediction of Landslide Risk in Coastal Areas Using GIS and Multicriteria Decision Making: A DeepLabV3+ Approach
title_full Deep Learning-Based Spatial Prediction of Landslide Risk in Coastal Areas Using GIS and Multicriteria Decision Making: A DeepLabV3+ Approach
title_fullStr Deep Learning-Based Spatial Prediction of Landslide Risk in Coastal Areas Using GIS and Multicriteria Decision Making: A DeepLabV3+ Approach
title_full_unstemmed Deep Learning-Based Spatial Prediction of Landslide Risk in Coastal Areas Using GIS and Multicriteria Decision Making: A DeepLabV3+ Approach
title_short Deep Learning-Based Spatial Prediction of Landslide Risk in Coastal Areas Using GIS and Multicriteria Decision Making: A DeepLabV3+ Approach
title_sort deep learning based spatial prediction of landslide risk in coastal areas using gis and multicriteria decision making a deeplabv3 approach
topic Coastal monitoring
DeepLabV3+
geographic information systems (GISs)
landslide susceptibility mapping
multisensor data fusion
remote sensing
url https://ieeexplore.ieee.org/document/11030867/
work_keys_str_mv AT huyongyan deeplearningbasedspatialpredictionoflandslideriskincoastalareasusinggisandmulticriteriadecisionmakingadeeplabv3approach
AT asadkhan deeplearningbasedspatialpredictionoflandslideriskincoastalareasusinggisandmulticriteriadecisionmakingadeeplabv3approach
AT ahsanjamil deeplearningbasedspatialpredictionoflandslideriskincoastalareasusinggisandmulticriteriadecisionmakingadeeplabv3approach
AT belkendilabdeldjalil deeplearningbasedspatialpredictionoflandslideriskincoastalareasusinggisandmulticriteriadecisionmakingadeeplabv3approach
AT taoufiksaidani deeplearningbasedspatialpredictionoflandslideriskincoastalareasusinggisandmulticriteriadecisionmakingadeeplabv3approach
AT nazihyrebouh deeplearningbasedspatialpredictionoflandslideriskincoastalareasusinggisandmulticriteriadecisionmakingadeeplabv3approach