isp-uv-es/multisenge-s2-taco

MultiSenGE (Sentinel-2 half)

This is a repackaging, not a new dataset. It is MultiSenGE (Sentinel-2 half) by LIVE / ICube, Universite de Strasbourg (Wenger et al.), Copernicus Sentinel-2 (ESA), via Theia, GeoGrandEst / IGN (OCSGE2 + BD TOPO, LULC reference), converted to TACO. Pixel values and labels are kept as released except where the description below says otherwise. All credit belongs to the original authors: if you use it, please cite them and follow their licence.

original dataset · paper · licence: CC-BY-4.0

Repackaged into TACO by the Image and Signal Processing Group (ISP), Universitat de València, within the ELLIOT project.

Citation

Please cite the original work:

@article{wenger2022multisenge,
  title   = {MultiSenGE: A Multimodal and Multitemporal Benchmark Dataset for Land Use/Land Cover Remote Sensing Applications},
  author  = {Wenger, Romain and Puissant, Anne and Weber, Jonathan and Idoumghar, Lhassane and Forestier, Germain},
  journal = {ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences},
  volume  = {V-3-2022},
  pages = {635--640},
  year    = {2022}
}

About the data

8,157 non-overlapping 256x256 patches at 10 m over the French Grand-Est region for 2020. Each patch carries a Sentinel-2 L2A series of 2-23 dates (10 bands, reflectance x 1e4) and a 14-class regional Land Use / Land Cover reference raster, plus the set of classes present in the patch as a scene-level multi-label.

8,157 samples · splits: test 855 · train 6,432 · validation 870 · tasks: semantic-segmentation, classification

Packaged as TACO v3.

Full description

Scope. The release's Sentinel-1 half is not carried. The authors' s1.tgz archive is truncated at the source (at about 20.7%), so none of the 8,157 patches can be converted from it, so this is the Sentinel-2 half rather than the triplet the paper describes.

Legend. The paper's level-3 typology. Its level-1 grouping is urban areas 1-5, agricultural 6-9, semi-natural 10-12, wetlands 13, water 14, and it maps onto ml:lc11. Class 0 never occurs, so nothing is ignored.

Splits. The release publishes none. This one is assigned per 25.6 km block by a stable hash: 6432 train, 870 val, 855 test. 402 patch pairs touch across a split boundary.

Series. Frames are sorted by date, which the release's own frame lists are not.

Getting started

git clone --recursive https://github.com/OscarPellicer/taco
pip install -e "taco/python[ml]"     # builds the reader: C++23, CMake, Ninja, pkg-config, libcurl >= 7.83, OpenSSL >= 3

Read it straight from the Hub:

from huggingface_hub import snapshot_download
from taco.ml import Dataset, plot_sample

path = snapshot_download("isp-uv-es/multisenge-s2-taco", repo_type="dataset")
ds = Dataset(path)                  # the repository is the container: parts + .tacocat
plot_sample(ds[0])

or from a local copy:

ds = Dataset("multisenge-s2")
sample = ds[0]                      # {slot name: SlotValue}, arrays decoded
sample["s2"].array.shape

Metadata without decoding anything:

import taco
taco.read("multisenge-s2/.tacocat")              # one Arrow table, levels joined

Samples

What a sample contains

role	slot	holds	modality	detail
input	s2	raster_series	multispectral	10 band(s), unit 1, scaled
target	lulc	mask	label_raster	15 classes
target	lulc_present	class_multihot		15 classes

Licence

CC-BY-4.0

Providers: LIVE / ICube, Universite de Strasbourg (Wenger et al.), Copernicus Sentinel-2 (ESA), via Theia, GeoGrandEst / IGN (OCSGE2 + BD TOPO, LULC reference)

Acknowledgements

TACO was designed by César Aybar and is specified at https://asterisk.coop/taco/spec/.

Built by Oscar Pellicer within the ELLIOT project at the Image and Signal Processing Group (ISP), Universitat de València.