isp-uv-es/chabud-taco

ChaBuD (burned area change detection)

This is a repackaging, not a new dataset. It is ChaBuD (burned area change detection) by Politecnico di Torino (Rege Cambrin, Colomba, Garza; CaBuAr / ChaBuD challenge), ESA Copernicus / 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-NC-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{cambrin2023cabuar,
  title   = {{CaBuAr}: California Burned Areas Dataset for Delineation},
  author  = {Rege Cambrin, Daniele and Colomba, Luca and Garza, Paolo},
  journal = {IEEE Geoscience and Remote Sensing Magazine},
  volume  = {11},
  number  = {3},
  pages   = {106--113},
  year    = {2023},
  doi     = {10.1109/MGRS.2023.3292467}
}

About the data

424 Sentinel-2 L2A before/after pairs of 512x512 at 10 m with twelve bands each, and a binary burned-area mask. Data of the ChaBuD ECML-PKDD 2023 Discovery Challenge.

424 samples · splits: test 68 · train 278 · validation 78 · tasks: change-detection

Packaged as TACO v3.

Full description

Splits. The challenge splits by fold: folds 1-4 are train and fold 0 is val. The challenge's held-out test fold, distributed later as a separate file with CaBuAr, is the test split (68 pairs); fold keeps each sample's fold.

Coverage. A subset of the release: entries that ship a post-fire image only cannot form a bitemporal pair and are excluded.

Storage. The pre/post pair is a two-frame raster series. Converted from the official HDF5.

Getting started

git clone https://github.com/OscarPellicer/taco
pip install -e "taco/python[ml]"     # builds the native reader: needs CMake and Ninja

Read it straight from the Hub:

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

path = hf_hub_download("isp-uv-es/chabud-taco", "chabud.zip", repo_type="dataset")
ds = Dataset(path)
plot_sample(ds[0])

or from a local copy:

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

Metadata without decoding anything:

import taco
taco.read("chabud.zip")              # one Arrow table, levels joined

Samples

What a sample contains

role	slot	holds	modality	detail
input	s2	raster_series	multispectral	12 band(s), unit 1, scaled
target	burned	mask	label_raster	2 classes

Licence

CC-BY-NC-4.0

Terms of use:

• Non-commercial use only (CC-BY-NC-4.0).
• Credit the original authors and the data sources listed below.

Required credits:

• CaBuAr / ChaBuD ECML-PKDD 2023 Discovery Challenge (Cambrin, Colomba, Garza), Politecnico di Torino
• Contains modified Copernicus Sentinel data

Providers: Politecnico di Torino (Rege Cambrin, Colomba, Garza; CaBuAr / ChaBuD challenge), ESA Copernicus / 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.