Further references#

Here, we provide a collection of relevant links provided by our research group:

  • The first BigEarthNet (S2) paper Sumbul et al. [6]

  • The BigEarthNet-MM publication + the recommended 19-class nomenclature Sumbul et al. [7]

The BigEarthNet Guide#

DOI

The BigEarthNet Guide documentation introduces the multi-modal BigEarthNet v1.0 dataset and makes it more accessible to others by providing an interactive dataset website.

Pretrained models#

Every repository includes code to re-run the training procedure. These models are all trained with the TensorFlow library.

  • Pretrained models trained on BigEarthNet-S2 with 43-classes

    • https://git.tu-berlin.de/rsim/BigEarthNet-S2_43-classes_models

  • Pretrained models trained on BigEarthNet-S2 with 19-classes

    • https://git.tu-berlin.de/rsim/BigEarthNet-S2_19-classes_models

  • Pretrained multi-modal models trained on BigEarthNet-S1 and BigEarthNet-S2 simultaneously

    • https://git.tu-berlin.de/rsim/BigEarthNet-MM_19-classes_models

  • Pretrained multi-modal models trained on refined BigEarthNet (ConfigILM was used for the official pretrained checkpoints)

    • https://huggingface.co/BIFOLD-BigEarthNetv2-0

BigEarthNet Tools#

  • BigEarthNet-S1 Tools

    • https://git.tu-berlin.de/rsim/BigEarthNet-S1_tools

    • Read GeoTIFF patches from BigEarthNet-S1

    • Script to extract names and download links of the Sentinel-1 Level-1C GRD tiles

      • Requires the BigEarthNet-S1 dataset on disk

  • BigEarthNet-S2 Tools

    • https://git.tu-berlin.de/rsim/BigEarthNet-S2_tools

    • Read GeoTIFF patches from BigEarthNet-S2

      • While skipping cloudy/snowy patches

    • Script to extract names and download links of the Sentinel-2 Level-1C tiles

      • Requires the BigEarthNet-S2 archive on disk

  • Code to read pairs of Sentinel-1 and Sentinel-2 patches

    • https://git.tu-berlin.de/rsim/BigEarthNet-MM_tools

  • Converter to create a high-throughput format for refined BigEarthNet

    • https://github.com/kai-tub/rico-hdl

Bibliography#

[1]

Kai Norman Clasen, Leonard Hackel, Tom Burgert, Gencer Sumbul, Begüm Demir, and Volker Markl. Reben: refined bigearthnet dataset for remote sensing image analysis. arXiv preprint arXiv:2407.03653, 2024.

[2]

Kun Li, George Vosselman, and Michael Ying Yang. Hrvqa: a visual question answering benchmark for high-resolution aerial images. arXiv preprint arXiv:2301.09460, 2023.

[3]

Sylvain Lobry, Begüm Demir, and Devis Tuia. Rsvqa meets bigearthnet: a new, large-scale, visual question answering dataset for remote sensing. In 2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS, 1218–1221. IEEE, 2021.

[4]

Sylvain Lobry, Diego Marcos, Jesse Murray, and Devis Tuia. Rsvqa: visual question answering for remote sensing data. IEEE Transactions on Geoscience and Remote Sensing, 58(12):8555–8566, 2020.

[5]

Mengye Ren, Ryan Kiros, and Richard Zemel. Exploring models and data for image question answering. Advances in neural information processing systems, 2015.

[6]

Gencer Sumbul, Marcela Charfuelan, Begüm Demir, and Volker Markl. Bigearthnet: A large-scale benchmark archive for remote sensing image understanding. In IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium. IEEE, July 2019. URL: https://doi.org/10.1109/igarss.2019.8900532, doi:10.1109/igarss.2019.8900532.

[7]

Gencer Sumbul, Arne de Wall, Tristan Kreuziger, Filipe Marcelino, Hugo Costa, Pedro Benevides, Mario Caetano, Begüm Demir, and Volker Markl. BigEarthNet-MM: A large-scale, multimodal, multilabel benchmark archive for remote sensing image classification and retrieval [Software and data sets]. IEEE Geosci. Remote Sens. Mag., 9(3):174–180, September 2021. URL: https://doi.org/10.1109/mgrs.2021.3089174, doi:10.1109/mgrs.2021.3089174.