Dual-stream feature fusion network for landslide detection
6 affiliations
- 1School of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology
- 2Division of Electronics and Electrical Engineering, Dongguk University
- 3School of Atmospheric Sciences, Nanjing University of Information Science and Technology
- 4Aerospace Information Research Institute, Henan Academy of Sciences
- 5Business School, Hohai University
- 6School of Ecology and Applied Meteorology, Nanjing University of Information Science and Technology
Abstract
Landslides are destructive natural hazards that threaten infrastructure, ecosystems, and human settlements, particularly in mountainous regions. Although deep learning has significantly improved landslide detection from remote sensing imagery, existing models often struggle with complex terrain conditions or require high computational resources. To address these challenges, this study proposes a Dual-Stream Feature Fusion Network (DSFF-Net) for accurate and efficient landslide segmentation. The proposed architecture integrates dual-stream feature extraction, hierarchical feature fusion, and an Enhanced Spatial Attention (ESA) module to effectively capture both spatial details and high-level semantic information. DSFF-Net was evaluated using two benchmark datasets, Bijie and Landslide4Sense. Experimental results show that the model achieves 96.34% precision, 97.83% recall, 97.08% F1-score, and 94.32% mIoU on the Bijie dataset, and 77.05% precision, 64.48% recall, 70.20% F1-score, and 72.36% mIoU on the Landslide4Sense dataset. Despite its compact architecture, DSFF-Net maintains low computational complexity with approximately 24.51 million parameters, demonstrating a strong balance between segmentation accuracy and computational efficiency. These results indicate that the proposed framework provides a reliable approach for automated landslide detection from remote sensing imagery and may support large-scale landslide inventory mapping and environmental monitoring applications. Future studies should further evaluate the model across diverse geographic regions and integrate multi-source geospatial data to enhance its generalization and operational applicability for landslide hazard monitoring.
Related
- Physics‐Informed Fusion Learning for Nowcasting Fault Detection of Hydrological Multi Sensor Network
- Mapping landslide change and volume with repeat airborne LiDAR after the May 2023 Emilia-Romagna event: slope-dependent detection limits and lowered areas
- Station spacing bounds the recovery of rainfall extremes by interpolation: an independent-network test with consequences for landslide thresholds
- Image-Based Asymmetric Feature Fusion Using EfficientNet-B0 and MobileNetV3Large for Multi-Class Fish Disease Classification
- A physics-constrained superpixel-based multi-source remote sensing spatiotemporal fusion network
BibTeX
@article{ullah2026dualstream,
title = {{Dual-stream feature fusion network for landslide detection}},
author = {Kaleem Ullah and Xie Tao and Tahir Mahmood and Sidra Syed and Irfan Ullah and Wilayat Shah and Muhammad Qasim and Rafi Riaz},
journal = {Scientific Reports},
year = {2026},
month = jul,
doi = {10.1038/s41598-026-59282-4},
url = {https://doi.org/10.1038/s41598-026-59282-4},
note = {CC BY},
}
RIS
TY - JOUR
TI - Dual-stream feature fusion network for landslide detection
AU - Kaleem Ullah
AU - Xie Tao
AU - Tahir Mahmood
AU - Sidra Syed
AU - Irfan Ullah
AU - Wilayat Shah
AU - Muhammad Qasim
AU - Rafi Riaz
PY - 2026
DA - 2026/07/01
JO - Scientific Reports
DO - 10.1038/s41598-026-59282-4
UR - https://doi.org/10.1038/s41598-026-59282-4
AB - Landslides are destructive natural hazards that threaten infrastructure, ecosystems, and human settlements, particularly in mountainous regions. Although deep learning has significantly improved landslide detection from remote sensing imagery, existing models often struggle with complex terrain conditions or require high computational resources. To address these challenges, this study proposes a Dual-Stream Feature Fusion Network (DSFF-Net) for accurate and efficient landslide segmentation. The proposed architecture integrates dual-stream feature extraction, hierarchical feature fusion, and an Enhanced Spatial Attention (ESA) module to effectively capture both spatial details and high-level semantic information. DSFF-Net was evaluated using two benchmark datasets, Bijie and Landslide4Sense. Experimental results show that the model achieves 96.34% precision, 97.83% recall, 97.08% F1-score, and 94.32% mIoU on the Bijie dataset, and 77.05% precision, 64.48% recall, 70.20% F1-score, and 72.36% mIoU on the Landslide4Sense dataset. Despite its compact architecture, DSFF-Net maintains low computational complexity with approximately 24.51 million parameters, demonstrating a strong balance between segmentation accuracy and computational efficiency. These results indicate that the proposed framework provides a reliable approach for automated landslide detection from remote sensing imagery and may support large-scale landslide inventory mapping and environmental monitoring applications. Future studies should further evaluate the model across diverse geographic regions and integrate multi-source geospatial data to enhance its generalization and operational applicability for landslide hazard monitoring.
SN - 2045-2322
ER -