A multi-scale factor feature fusion modeling method for landslide susceptibility mapping

Guofang Wang; Jun Cao; Jianwen Sun; Yifan Wang; Hao Geng; Shouhong Ye; Guofang Wang; Jun Cao; Jianwen Sun; Yifan Wang; Hao Geng; Shouhong Ye

doi:10.3934/geosci.2026013

AIMS Geosciences

2026, Volume 12, Issue 2: 335-359. doi: 10.3934/geosci.2026013

Previous Article Next Article

Research article

A multi-scale factor feature fusion modeling method for landslide susceptibility mapping

Electric Power Research Institute, Yunnan Power Grid Company ltd., 650217, China

Received: 29 August 2025 Revised: 17 December 2025 Accepted: 19 January 2026 Published: 20 March 2026

Landslide susceptibility mapping (LSM) is a critical technique for geological hazard risk assessment, aiming to quantitatively evaluate and spatially delineate potentially susceptible areas. Recently, convolutional neural networks (CNNs) and Transformer architectures have been widely applied in remote sensing and geospatial contexts, significantly enhancing automatic feature extraction capabilities. However, single-model approaches often struggle to capture fine-grained local characteristics and long-range spatial dependencies in landslide-prone regions, limiting their representation of multi-scale and irregular spatial structures. To address this challenge, we propose a progressive local-to-global synergy network (PLGS-Net), which integrates the local spatial perception of CNNs with the global dependency modeling of Transformers to achieve hierarchical, collaborative representation of landslide susceptibility features. Empirical experiments conducted in a typical landslide-prone region of Ya'an, Sichuan Province, demonstrated that PLGS-Net outperforms conventional CNNs, pure Transformers, and other mainstream deep learning models in classification accuracy, generalization ability, and regional adaptability, providing an efficient and practical approach for landslide susceptibility assessment in complex terrains.
- landslide susceptibility mapping,
- deep learning,
- PLGS-Net
Citation: Guofang Wang, Jun Cao, Jianwen Sun, Yifan Wang, Hao Geng, Shouhong Ye. A multi-scale factor feature fusion modeling method for landslide susceptibility mapping[J]. AIMS Geosciences, 2026, 12(2): 335-359. doi: 10.3934/geosci.2026013

Related Papers:

Abstract

Landslide susceptibility mapping (LSM) is a critical technique for geological hazard risk assessment, aiming to quantitatively evaluate and spatially delineate potentially susceptible areas. Recently, convolutional neural networks (CNNs) and Transformer architectures have been widely applied in remote sensing and geospatial contexts, significantly enhancing automatic feature extraction capabilities. However, single-model approaches often struggle to capture fine-grained local characteristics and long-range spatial dependencies in landslide-prone regions, limiting their representation of multi-scale and irregular spatial structures. To address this challenge, we propose a progressive local-to-global synergy network (PLGS-Net), which integrates the local spatial perception of CNNs with the global dependency modeling of Transformers to achieve hierarchical, collaborative representation of landslide susceptibility features. Empirical experiments conducted in a typical landslide-prone region of Ya'an, Sichuan Province, demonstrated that PLGS-Net outperforms conventional CNNs, pure Transformers, and other mainstream deep learning models in classification accuracy, generalization ability, and regional adaptability, providing an efficient and practical approach for landslide susceptibility assessment in complex terrains.

References

[1]	Gariano SL, Guzzetti F (2016) Landslides in a changing climate. Earth-Sci Rev 162: 227–252. https://doi.org/10.1016/j.earscirev.2016.08.011 doi: 10.1016/j.earscirev.2016.08.011
[2]	Froude MJ, Petley DN (2018) Global fatal landslide occurrence from 2004 to 2016. Nat Hazards Earth Syst Sci 18: 2161–2181. https://doi.org/10.5194/nhess-18-2161-2018 doi: 10.5194/nhess-18-2161-2018
[3]	Ngo PTT, Panahi M, Khosravi K, et al. (2021) Evaluation of deep learning algorithms for national scale landslide susceptibility mapping of Iran. Geosci Front 12: 505–519. https://doi.org/10.1016/j.gsf.2020.06.013 doi: 10.1016/j.gsf.2020.06.013
[4]	Fang Y, Gan C, Cao W (2023) Undersampling Ensemble and Deep Learning-Based Landslide Susceptibility Mapping Method for Geological Hazard Warning. China Automation Congress (CAC), IEEE. https://doi.org/10.1109/CAC59555.2023.10450744
[5]	Zhao Y, Wang R, Jiang Y, et al. (2019) GIS-based logistic regression for rainfall-induced landslide susceptibility mapping under different grid sizes in Yueqing, Southeastern China. Eng Geol 259: 105147. https://doi.org/10.1016/j.enggeo.2019.105147 doi: 10.1016/j.enggeo.2019.105147
[6]	Guo Y, Dou J, Xiang Z, et al. (2024) Susceptibility evaluation of Wenchuan coseismic landslides by gradient boosting decision tree and random forest based on optimal negative sample sampling strategies. Bull Geol Sci Technol 43: 251–265. https://doi.org/10.19509/j.cnki.dzkq.tb20230037 doi: 10.19509/j.cnki.dzkq.tb20230037
[7]	Sarkar P, Mondal M, Sarkar A, et al. (2025) Landslide susceptibility assessment for the Darjeeling Toy Train route: a GIS and machine learning approach. Stoch Environ Res Risk Assess 39: 613–637. https://doi.org/10.1007/s00477-024-02885-y doi: 10.1007/s00477-024-02885-y
[8]	Sun D, Chen D, Mi C, et al. (2023) Evaluation of landslide susceptibility in the gentle hill-valley areas based on the interpretable random forest-recursive feature elimination model. J Geomech 29: 202–219. https://doi.org/10.12090/j.issn.1006-6616.2022128 doi: 10.12090/j.issn.1006-6616.2022128
[9]	Liu S, Wang L, Zhang W, et al. (2023) A comprehensive review of machine learning-based methods in landslide susceptibility mapping. Geol J 58: 2283–2301. https://doi.org/10.1002/gj.4666 doi: 10.1002/gj.4666
[10]	Juyal A, Sharma S (2021) A study of landslide susceptibility mapping using machine learning approach. 2021 Third International Conference on Intelligent Communication Technologies and Virtual Mobile Networks (ICICV). IEEE, 1523–1528. https://doi.org/10.1109/ICICV50876.2021.9388379
[11]	Merghadi A, Yunus AP, Dou J, et al. (2020) Machine learning methods for landslide susceptibility studies: A comparative overview of algorithm performance. Earth-Sci Rev 207: 103225. https://doi.org/10.1016/j.earscirev.2020.103225 doi: 10.1016/j.earscirev.2020.103225
[12]	Wei R, Ye C, Sui T, et al. (2022) Combining spatial response features and machine learning classifiers for landslide susceptibility mapping. Int J Appl Earth Obs Geoinf 107: 102681. https://doi.org/10.1016/j.jag.2022.102681 doi: 10.1016/j.jag.2022.102681
[13]	Pham BT, Prakash I, Dou J, et al. (2020) A novel hybrid approach of landslide susceptibility modelling using rotation forest ensemble and different base classifiers. Geocarto Int 35: 1267–1292. https://doi.org/10.1080/10106049.2018.1559885 doi: 10.1080/10106049.2018.1559885
[14]	Yesilnacar E, Topal T (2005) Landslide susceptibility mapping: a comparison of logistic regression and neural networks methods in a medium scale study, Hendek region (Turkey). Eng Geol 79: 251–266. https://doi.org/10.1016/j.enggeo.2005.02.002 doi: 10.1016/j.enggeo.2005.02.002
[15]	Budimir MEA, Atkinson PM, Lewis HG (2015) A systematic review of landslide probability mapping using logistic regression. Landslides 12: 419–436. https://doi.org/10.1007/s10346-014-0550-5 doi: 10.1007/s10346-014-0550-5
[16]	Huang F, Chen B, Mao D, et al. (2023) Landslide susceptibility prediction modeling and interpretability based on self-screening deep learning model. Earth Sci 48: 1696–1710.
[17]	Feng Y, Xu H, Jiang J, et al. (2022) ICIF-Net: Intra-scale cross-interaction and inter-scale feature fusion network for bitemporal remote sensing images change detection. IEEE Trans Geosci Remote Sens 60: 1–13. https://doi.org/10.1109/TGRS.2022.3168331 doi: 10.1109/TGRS.2022.3168331
[18]	Ding L, Hong D, Zhao M, et al. (2025) A survey of sample-efficient deep learning for change detection in remote sensing: Tasks, strategies, and challenges. IEEE Geosci Remote Sens Mag 13: 164–189. https://doi.org/10.1109/MGRS.2025.3533605 doi: 10.1109/MGRS.2025.3533605
[19]	Youssef AM, Pradhan B, Dikshit A, et al. (2022) Landslide susceptibility mapping using CNN-1D and 2D deep learning algorithms: comparison of their performance at Asir Region, KSA. Bull Eng Geol Environ 81: 165. https://doi.org/10.1007/s10064-022-02657-4 doi: 10.1007/s10064-022-02657-4
[20]	Gao Z, Hu A, Chen B, et al. (2024) A hierarchical geometry-to-semantic fusion GNN framework for earth surface anomalies detection. Advances in Brain Inspired Cognitive Systems. BICS 2023. Lecture Notes in Computer Science, 14374. Springer, Singapore. https://doi.org/10.1007/978-981-97-1417-9_6
[21]	Wang Y, Fang Z, Hong H (2019) Comparison of convolutional neural networks for landslide susceptibility mapping in Yanshan County, China. Sci Total Environ 666: 975–993. https://doi.org/10.1016/j.scitotenv.2019.02.263 doi: 10.1016/j.scitotenv.2019.02.263
[22]	Sameen MI, Pradhan B, Lee S (2020) Application of convolutional neural networks featuring Bayesian optimization for landslide susceptibility assessment. Catena 186: 104249. https://doi.org/10.1016/j.catena.2019.104249 doi: 10.1016/j.catena.2019.104249
[23]	Huang F, Zhang J, Zhou C, et al. (2020) A deep learning algorithm using a fully connected sparse autoencoder neural network for landslide susceptibility prediction. Landslides 17: 217–229. https://doi.org/10.1007/s10346-019-01274-9 doi: 10.1007/s10346-019-01274-9
[24]	Chen C, Fan L (2022) CNN-LSTM-attention deep learning model for mapping landslide susceptibility in Kerala, India. ISPRS Ann Photogramm Remote Sens Spatial Inf Sci 10: 25–30. https://doi.org/10.5194/isprs-annals-X-3-W1-2022-25-2022 doi: 10.5194/isprs-annals-X-3-W1-2022-25-2022
[25]	Zhang B, Tang J, Huan Y, et al. (2024) Multi-scale convolutional neural networks (CNNs) for landslide inventory mapping from remote sensing imagery and landslide susceptibility mapping (LSM). Geomat Nat Hazards Risk 15: 2383309. https://doi.org/10.1080/19475705.2024.2383309 doi: 10.1080/19475705.2024.2383309
[26]	Huang W, Ding M, Li Z, et al. (2023) Landslide susceptibility mapping and dynamic response along the Sichuan-Tibet transportation corridor using deep learning algorithms. Catena 222: 106866. https://doi.org/10.1016/j.catena.2022.106866 doi: 10.1016/j.catena.2022.106866
[27]	Wang D, Yang RH, Wang X, et al. (2023) Evaluation of deep learning algorithms for landslide susceptibility mapping in an alpine-gorge area: a case study in Jiuzhaigou County. J Mt Sci 20: 484–500. https://doi.org/10.1007/s11629-022-7326-5 doi: 10.1007/s11629-022-7326-5
[28]	Jamali A, Roy SK, Ghamisi P (2023) WetMapFormer: A unified deep CNN and vision transformer for complex wetland mapping. Int J Appl Earth Obs Geoinf 120: 103333. https://doi.org/10.1016/j.jag.2023.103333 doi: 10.1016/j.jag.2023.103333
[29]	Bao S, Liu J, Wang L, et al. (2022) Application of transformer models to landslide susceptibility mapping. Sensors 22: 9104. https://doi.org/10.3390/s22239104 doi: 10.3390/s22239104
[30]	Zhao Z, Chen T, Dou J, et al. (2024) Landslide susceptibility mapping considering landslide local-global features based on CNN and transformer. IEEE J Sel Top Appl Earth Obs Remote Sens 17: 7475–7489. https://doi.org/10.1109/JSTARS.2024.3379350 doi: 10.1109/JSTARS.2024.3379350
[31]	Zhu Q, Zhang M, Ding Y, et al. (2021) Fuzzy Logic Approach for Regional Landslide Susceptibility Analysis Constrained by Spatial Characteristics of Environmental Factors. Geomatics Inf Sci Wuhan Univ 46: 1431–1440.
[32]	Bian J, Li X, Hu K (2018) Study on distribution characteristics and dynamic evolution of mountain hazards in Hengduan Mountains area. J Eng Geol 26: 6–13. https://doi.org/10.13544/j.cnki.jeg.2018030 doi: 10.13544/j.cnki.jeg.2018030
[33]	Chang Z, Huang F, Jiang S, et al. (2023) Slope unit extraction and landslide susceptibility prediction using multi-scale segmentation method. Adv Eng Sci 55: 184–195.
[34]	Data Center for Resources and Environmental Sciences, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences. Spatial Distribution Data of Geological Disaster Points [DB/OL]. Beijing, 2025. Available from: http://www.resdc.cn/.
[35]	Guo Z, Wang H, He J, et al. (2025) PSLSA v2. 0: An automatic Python package integrating machine learning models for regional landslide susceptibility assessment. Environ Model Softw 186: 106367. https://doi.org/10.1016/j.envsoft.2025.106367
[36]	Wang Y, Sun D, Wen H, H. et al. (2020) Comparison of random forest model and frequency ratio model for landslide susceptibility mapping (LSM) in Yunyang County (Chongqing, China). Int J Environ Res Public Health 17: 4206. https://doi.org/10.3390/ijerph17124206 doi: 10.3390/ijerph17124206
[37]	Aalianvari A, Jahanmiri S (2025) A Comparative Study of TabNet and Classical Machine Learning Models for Landslide Prediction [Preprint]. bioRxiv: https://doi.org/10.21203/rs.3.rs-7694229/v1
[38]	Aslam B, Zafar A, Khalil U (2023) Comparative analysis of multiple conventional neural networks for landslide susceptibility mapping. Nat Hazards 115: 673–707. https://doi.org/10.1007/s11069-022-05570-x doi: 10.1007/s11069-022-05570-x
[39]	Lv L, Chen T, Dou J, et al. (2022) A hybrid ensemble-based deep-learning framework for landslide susceptibility mapping. Int J Appl Earth Obs Geoinf 108: 102713. https://doi.org/10.1016/j.jag.2022.102713 doi: 10.1016/j.jag.2022.102713
[40]	Ge Y, Liu G, Tang H, et al. (2023) Comparative analysis of five convolutional neural networks for landslide susceptibility assessment. Bull Eng Geol Environ 82: 377. https://doi.org/10.1007/s10064-023-03408-9 doi: 10.1007/s10064-023-03408-9
[41]	Chen T, Wang Q, Zhao Z, et al. (2024) LCFSTE: Landslide conditioning factors and swin transformer ensemble for landslide susceptibility assessment. IEEE J Sel Top Appl Earth Obs Remote Sens 17: 6444–6454. https://doi.org/10.1109/JSTARS.2024.3373029 doi: 10.1109/JSTARS.2024.3373029
[42]	Zhu Y, Liu S, Yin K, et al. (2025) Impact of negative sampling strategies on landslide susceptibility assessment. Adv Space Res 76: 592–613. https://doi.org/10.1016/j.asr.2025.04.070 doi: 10.1016/j.asr.2025.04.070
[43]	Zhu Y, Yin K, Yang H, et al. (2026) Temporal validity of landslide inventories in hazard mapping: insights from Hubei Province, China. Landslides 23: 417–432. https://doi.org/10.1007/s10346-025-02651-3 doi: 10.1007/s10346-025-02651-3

Reader Comments

Your name:*

Email:*
© 2026 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)