README.md

---
language:
- en
license: cc-by-4.0
task_categories:
- object-detection
pretty_name: Global Wheat Head
tags:
- agriculture
- biology
dataset_info:
  features:
  - name: image
    dtype: image
  - name: domain
    dtype: string
  - name: country
    dtype: string
  - name: location
    dtype: string
  - name: development_stage
    dtype: string
  - name: objects
    struct:
    - name: boxes
      sequence:
        sequence: int64
    - name: categories
      sequence: int64
  splits:
  - name: train
    num_bytes: 701105106.93
    num_examples: 3655
  - name: validation
    num_bytes: 264366740.324
    num_examples: 1476
  - name: test
    num_bytes: 301053063.17
    num_examples: 1381
  download_size: 1260938177
  dataset_size: 1266524910.424
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---


# Dataset Card for "Global Wheat Head Dataset 2021" 😊

If you want any update on the Global Wheat Dataset Community, go on https://www.global-wheat.com/

## Table of Contents
- [Dataset Description](#dataset-description)
- [Dataset Composition](#dataset-composition)
- [Usage](#usage)
- [Citation](#citation)
- [Acknowledgements](#acknowledgements)

## Dataset Description

- **Creators**: Etienne David and others
- **Published**: July 12, 2021 | Version 1.0
- **Availability**: [Zenodo Link](https://doi.org/10.5281/zenodo.5092309)
- **Keywords**: Deep Learning, Wheat Counting, Plant Phenotyping

### Introduction
Wheat is essential for a large part of humanity. The "Global Wheat Head Dataset 2021" aims to support the development of deep learning models for wheat head detection. This dataset addresses challenges like overlapping plants and varying conditions across global wheat fields. It's a step towards automating plant phenotyping and enhancing agricultural practices. 🌾

### Dataset Composition
- **Images**: Over 6000, Resolution - 1024x1024 pixels
- **Annotations**: 300k+ unique wheat heads with bounding boxes
- **Geographic Coverage**: Images from 11 countries
- **Domains**: Various, including sensor types and locations
- **Splits**: Training (Europe & Canada), Test (Other regions)

## Dataset Composition

### Files and Structure
- **Images**: Folder containing all images (`.png`)
- **CSV Files**: `competition_train.csv`, `competition_val.csv`, `competition_test.csv` for different dataset splits
- **Metadata**: `Metadata.csv` with additional details

### Labels
- **Format**: CSV with columns - image_name, BoxesString, domain
- **BoxesString**: `[x_min,y_min, x_max,y_max]` format for bounding boxes
- **Domain**: Specifies the image domain

## Usage

### Tutorials and Resources
- Tutorials available at [AIcrowd Challenge Page](https://www.aicrowd.com/challenges/global-wheat-challenge-2021)

### License
- **Type**: Creative Commons Attribution 4.0 International (cc-by-4.0)
- **Details**: Free to use with attribution

## Citation

If you use this dataset in your research, please cite the following:

```bibtex
@article{david2020global,
  title={Global Wheat Head Detection (GWHD) dataset: a large and diverse dataset of high-resolution RGB-labelled images to develop and benchmark wheat head detection methods},
  author={David, Etienne and others},
  journal={Plant Phenomics},
  volume={2020},
  year={2020},
  publisher={Science Partner Journal}
}
@misc{david2021global,
  title={Global Wheat Head Dataset 2021: more diversity to improve the benchmarking of wheat head localization methods},
  author={Etienne David and others},
  year={2021},
  eprint={2105.07660},
  archivePrefix={arXiv},
  primaryClass={cs.CV}
}
```

## Acknowledgements

Special thanks to all the contributors, researchers, and institutions that played a pivotal role in the creation of this dataset. Your efforts are helping to advance the field of agricultural sciences and technology. 👏