A globally distributed dataset of coseismic landslide mapping via multi-source high-resolution remote sensing images
Citation
Fang, C., Fan, X., Wang, X., Nava, L., Zhong, H., Dong, X., Qi, J., Catani, F. (2024). A globally distributed dataset of coseismic landslide mapping via multi-source high-resolution remote sensing images. Earth System Science Data, 16(10): 4817-4842. Link to paper
Abstract
This work introduces the Globally Distributed Coseismic Landslide Dataset (GDCLD), integrating multi-source remote sensing images, including PlanetScope, Gaofen-6, Map World, and uncrewed aerial vehicle (UAV) data across nine earthquake events worldwide. The dataset enables evaluation of semantic segmentation algorithms, with the GDCLD-SegFormer model demonstrating superior performance. Testing on four independent regions confirmed the model's effectiveness. Notably, models demonstrated excellent performance when applied to rainfall-triggered landslides, establishing this dataset's broad applicability for landslide detection following unexpected events globally.