A dual-encoder U-Net for landslide detection using Sentinel-2 and DEM data
Citation
Lu, W., Hu, Y., Zhang, Z., Liu, J. (2023). A dual-encoder U-Net for landslide detection using Sentinel-2 and DEM data. Landslides, 20: 1975-1987. Link to paper
Abstract
This study develops a new deep learning-based method to detect landslides using medium-resolution imagery and digital elevation model (DEM) data which are free-access and covered globally. The Sentinel-2 imagery and NASADEM data are freely downloaded from Google Earth Engine platform. A workflow for constructing the landslide dataset was developed, then a semantic segmentation model was designed to learn deep features and generate per-pixel landslide predictions. The method achieved a best F1 score of 79.24%, outperforming SegNet, U-Net, and Attention U-Net, which are models commonly used in semantic segmentation-based landslide detection. The proposed method may have application potential in disaster risk assessment and post-disaster reconstruction and provide a technical reference for large-scale landslide mapping.