Skip to content

tselea/agrisen-cog

Repository files navigation

AgriSen-COG: a Large-Scale Dataset for Crop Detection

Table Of Contents

  1. Introduction
  2. Dataset
  3. Hands-on Tutorials
  4. Project's Overview
  5. Model Weights
  6. Acknowledgements

Introduction

Code for AgriSen-COG dataset. This repository is going to be public once the dataset is published.

This repository introduces the AgriSen-COG Dataset, a multicountry, multitemporal large-scale Sentinel-2 Benchmark Dataset for Crop Detection.

Dataset Repositories

Each repository contains the dataset, including the following - Austria (2019, 2020), Belgium (2019, 2020), Catalonia(2019, 2020), Denmark (2019, 2020), Netherlands (2019, 2020).

Dropbox

Link: https://www.dropbox.com/sh/5bc55skio0o5xd7/AAAQVG3ZmVGFNvPiltQ9Esqma?dl=0

  • AgriSen-COG/:
    • intermediate_outputs/: Contains the intermediate outputs of dataset preparation.
      • output1_1_original_lpis/: Contains the original LPIS.
      • output1_2_lpis_gpkg/: Contains the original LPIS as GPKG.
      • output1_2_lpis_parquet/: Contains the original LPIS as partioned Parquet files.
      • output1_3_lpis_en/: Contains the english version of each LPIS as partitioned Parquet.
      • output1_4_lpis_icc/: Contains the ICC code mapped version of each english LPIS as partitioned Parquet.

Zenodo

Link: https://doi.org/10.5281/zenodo.7892012

  • output1_1_original_lpis.zip: Contains the original LPIS.
  • output1_2_lpis_gpkg.zip: Contains the original LPIS as GPKG.
  • output1_2_lpis_parquet.zip: Contains the original LPIS as partioned Parquet files.
  • output1_3_lpis_en.zip: Contains the english version of each LPIS as partitioned Parquet.
  • output1_4_lpis_icc.zip: Contains the ICC code mapped version of each english LPIS as partitioned Parquet.

Minio S3 Bucket

Endpoint:

Bucket name: agrisen-cog-v1

Set anonymous access.

  • agrisen-cog-v1/:
    • LPIS_processing/:
      • original_files/: Contains the original LPIS as partioned Parquet files.
      • en_files/: Contains the english version of each LPIS as partitioned Parquet.
      • icc_en_files/: Contains the ICC code mapped version of each english LPIS as partitioned Parquet.

Hands-on Tutorials

In the notebooks folder we provide an overview of the following tasks:

  • OriginalLPISView.ipynb - View the original LPIS data as gpkg (local files) and Parquet (local and S3).
  • EnLPISView.ipynb - View the english and standardized LPIS data as partioned Parquet (S3 reading example).
  • ICCLPISView - View the ICC code mapped version of each english LPIS as partitioned Parquet (S3 reading example).

Project's Overview

Each package/directory contains usage examples.

  • lpis_files: Directory with multiple LPIS descriptions.
  • lpis_processing: Package for the LPIS data processing.

If you use our code, please cite:

Selea, Teodora. "AgriSen-COG, a Multicountry, Multitemporal Large-Scale Sentinel-2 Benchmark Dataset for Crop Mapping Using Deep Learning." Remote Sensing 15.12 (2023): 2980.

@article{selea2023agrisen,
  title={AgriSen-COG, a Multicountry, Multitemporal Large-Scale Sentinel-2 Benchmark Dataset for Crop Mapping Using Deep Learning},
  author={Selea, Teodora},
  journal={Remote Sensing},
  volume={15},
  number={12},
  pages={2980},
  year={2023},
  publisher={MDPI}
}

Selea, T. (2023). AgriSen-COG.LPIS.GT. [Data set]. Zenodo. https://doi.org/10.5281/zenodo.7892012

@dataset{selea2023agrisen,
  author       = {Selea, T.},
  title        = {AgriSen-COG.LPIS.GT.},
  year         = {2023},
  publisher    = {Zenodo},
  doi          = {10.5281/zenodo.7892012},
  url          = {https://doi.org/10.5281/zenodo.7892012},
}

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published