Sen12Landslides (S12LS)
Keywords: Landslides, Sentinel-1, Sentinel-2, DEM, SITS
Sen12Landslides is a large-scale, multi-modal dataset featuring 75,000 annotations from Sentinel-1/2 timeseries and DEM data designed to advance spatio-temporal landslide detection and geohazard monitoring. By providing precise pre- and post-event imagery across 15 global regions, it enables deep learning models to capture the complex temporal dynamics of landslide behavior.
Publications
A Spatio-Temporal Dataset for Satellite-Based Landslide Detection
Höhn P, Heidler K, Behling R, Zhu X - Scientific Data - 2025
Sen12Landslides
Paul Höhn, Konrad Heidler, Robert Behling, Xiao Xiang Zhu - Hugging Face - 2025