Copernicus-Bench / README.md
wangyi111's picture
Add link to code repository (#4)
a287ab1 verified
metadata
task_categories:
  - image-classification
  - image-segmentation
pretty_name: Copernicus-Bench
tags:
  - earth-observation
  - remote-sensing
  - benchmark
  - foundation-model

Dataset Card for Copernicus-Bench

A hierarchical ML benchmark for Copernicus Sentinels, with 15 datasets spread into three task levels covering all major Sentinel missions (S1,2,3,5P).

(Officially named "Copernicus-Bench", initially named "SentinelBench")

Dataset Details

Level Name Modality Task # Images Image Size # Classes Source License
L1 Cloud-S2 S2 TOA segmentation (cloud) 1699/567/551 512x512x13 4 CloudSEN12 CC 0 1.0
L1 Cloud-S3 S3 OLCI segmentation (cloud) 1197/399/399 256x256x21 5 new CC BY 4.0
L2 EuroSAT-S1 S1 GRD classification (LULC) 16200/5400/5400 64x64x2 10 EuroSAT-SAR CC BY 4.0
L2 EuroSAT-S2 S2 TOA classification (LULC) 16200/5400/5400 64x64x13 10 EuroSAT MIT
L2 BigEarthNet-S1 S1 GRD classification (LULC) 11894/6117/5991 120x120x12 19 BigEarthNet v2.0 CDLA-Permissive-1.0
L2 BigEarthNet-S2 S2 SR classification (LULC) 11894/6117/5991 120x120x12 19 BigEarthNet v2.0 CDLA-Permissive-1.0
L2 LC100Cls-S3 S3 OLCI classification (LULC) 5181/1727/1727* 96x96x21 23 new CC BY 4.0
L2 DFC2020-S1 S1 GRD segmentation (LULC) 3156/986/986 256x256x13 10 DFC2020 CC BY 4.0
L2 DFC2020-S2 S2 TOA segmentation (LULC) 3156/986/986 256x256x13 10 DFC2020 CC BY 4.0
L2 LC100Seg-S3 S3 OLCI segmentation (LULC) 5181/1727/1727* 96x96x21 (288x288) 23 new CC BY 4.0
L3 Flood-S1 S1 GRD change detection (flood) 3000/1000/1000 224x224x2 3 Kuro Siwo MIT
L3 LCZ-S2 S2 TOA classification (local climate zone) 15000/5000/5000 32x32x10 17 So2Sat LCZ42 CC BY 4.0
L3 Biomass-S3 S3 OLCI regression (biomass) 3000/1000/1000* 96x96x21 (288x288) 1 new CC BY 4.0
L3 AQ-NO2-S5P S5P NO2 regression (air quality) 1480/493/494* 56x56x1 1 new CC BY 4.0
L3 AQ-O3-S5P S5P O3 regression (air quality) 1480/493/494* 56x56x1 1 new CC BY 4.0

L1: preprocessing, L2: base applications, L3: specialized applications. *: time series available.

Related Sources

License

See individual datasets.

Citation

@misc{wang2025unifiedcopernicusfoundationmodel,
      title={Towards a Unified Copernicus Foundation Model for Earth Vision}, 
      author={Yi Wang and Zhitong Xiong and Chenying Liu and Adam J. Stewart and Thomas Dujardin and Nikolaos Ioannis Bountos and Angelos Zavras and Franziska Gerken and Ioannis Papoutsis and Laura Leal-Taixé and Xiao Xiang Zhu},
      year={2025},
      eprint={2503.11849},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2503.11849}, 
}