isp-uv-es/biomassters-taco

BioMassters

This is a repackaging, not a new dataset. It is BioMassters by Nascetti et al. / DrivenData (BioMassters), Finnish Forest Centre (LiDAR reference), ESA Copernicus / Sentinel-1, Sentinel-2, 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:

@inproceedings{nascetti2023biomassters,
title     = {BioMassters: A Benchmark Dataset for Forest Biomass Estimation using Multi-modal Satellite Time-series},
author    = {Nascetti, Andrea and Yadav, Ritu and Brodt, Kirill and Qu, Qixun and Fan, Hongwei and Shendryk, Yuri and Shah, Isha and Chung, Christine},
booktitle = {NeurIPS Datasets and Benchmarks},
year      = {2023}
}

About the data

Above-ground biomass regression over 11462 Finnish forest patches of 256x256 at 10 m.

11,462 samples · splits: test 2,773 · train 7,821 · validation 868 · tasks: regression

Packaged as TACO v3.

Full description

Contents. Each patch carries a full year of monthly Sentinel-1 (four channels: ascending VV/VH and descending VV/VH, dB, twelve months for every chip) and monthly Sentinel-2 L2A (ten reflectance bands plus the Sen2Cor cloud probability, 5 to 12 months present), against a dense LiDAR-derived above-ground biomass map in Mg/ha.

• The Sentinel-2 series is ragged and gappy, so each frame records which calendar month it is rather than only its position.
• The released GeoTIFFs carry no CRS or geotransform.

Splits. The test split is official; the train/val cut is a community one and is flagged as such.

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/biomassters-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("biomassters")
sample = ds[0]                      # {slot name: SlotValue}, arrays decoded
sample["s1"].array.shape

Metadata without decoding anything:

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

Samples

What a sample contains

role	slot	holds	modality	detail
input	s1	raster_series	sar	4 band(s), unit dB, physical
input	s2	raster_series	multispectral	11 band(s), unit 1, scaled
target	above_ground_biomass	raster	biophysical	1 band(s), unit Mg ha-1, physical

Licence

CC-BY-4.0

Providers: Nascetti et al. / DrivenData (BioMassters), Finnish Forest Centre (LiDAR reference), ESA Copernicus / Sentinel-1, Sentinel-2

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.