Landslide4Sense: Multi-Sensor Landslide Detection Benchmark

Provider
Institute of Advanced Research in Artificial Intelligence (IARAI); hosted by IBM-NASA Geospatial
License
CC-BY-4.0 (source)
Size
9,695 files
Format
HDF5 (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/ibm-nasa-geospatial/Landslide4sense)
Access
Free and public. No account with the data provider is needed.
Category
Earth Observation & Geospatial
Source
https://doi.org/10.1109/JSTARS.2022.3220845

About Landslide4Sense: Multi-Sensor Landslide Detection Benchmark

Benchmark for landslide detection and segmentation from multi-source satellite data, released by IARAI with the Landslide4Sense 2022 competition and drawn from landslide-affected regions worldwide between 2015 and 2021. Each 128 x 128 image patch is a 14-band composite at about 10 m: Sentinel-2 multispectral bands B1 to B12, plus slope and a digital elevation model from ALOS PALSAR, with a pixel-wise binary mask (landslide or not). Splits hold 3,799 training, 245 validation and 800 test patches. This copy stores each patch and mask as HDF5; the listing (measured 2026-11-14) shows images/ and annotations/ folders with train, validation and test subfolders (image_N.h5, mask_N.h5), and the repository uses about 12.1 GB. Read with h5py and numpy. The IARAI Zenodo record releases the reference data under CC-BY-4.0; cite Ghorbanzadeh et al., IEEE JSTARS 2022.

How to load Landslide4Sense: Multi-Sensor Landslide Detection Benchmark in Python

from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="ibm-nasa-geospatial/Landslide4sense", filename="annotations/test/mask_1.h5", repo_type="dataset")
print(path)

Data structure

File types in sampled files: HDF5

Files

Example file: annotations/test/mask_1.h5

Analyze this dataset without downloading it

On Scigantic, Landslide4Sense: Multi-Sensor Landslide Detection Benchmark opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Earth Observation & Geospatial datasets