---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: objects
    struct:
    - name: bbox
      list:
        list: float64
    - name: categories
      list:
        class_label:
          names:
            '0': ragweed
            '1': waterhemp
            '2': horseweed
            '3': redrootpigweed
            '4': kochia
  splits:
  - name: train
    num_bytes: 4554904299
    num_examples: 551
  download_size: 4554986069
  dataset_size: 4554904299
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: cc-by-4.0
task_categories:
- object-detection
size_categories:
- 1K<n<10K
---
# ImageWeeds Aerial Weed Detection

A dataset for object detection of weeds within crop fields. The dataset contains 551 images with 2,388 bounding box annotations across 5 categories.

This dataset is indexed on https://project-agml.github.io/ as part of the AgML python library.

## Citation

```bibtex
@article{rai2023multi,
  title={Multi-format open-source weed image dataset for real-time weed identification in precision agriculture},
  author={Rai, Nitin and Mahecha, Maria Villamil and Christensen, Annika and Quanbeck, Jamison and Zhang, Yu and Howatt, Kirk and Ostlie, Michael and Sun, Xin},
  journal={Data in Brief},
  volume={51},
  pages={109691},
  year={2023},
  publisher={Elsevier}
}
```

Rai, Nitin; Villamil Mahecha, Maria; Christensen, Annika; Quanbeck, Jamison; Howatt, Kirk; Ostlie, Michael; Zhang, Yu; Sun, Xin (2023), “ImageWeeds: An Image dataset consisting of weeds in multiple formats to advance computer vision algorithms for real-time weed identification and spot spraying application”, Mendeley Data, V2, doi: 10.17632/8kjcztbjz2.2

*This dataset was reformatted from its original format to match HuggingFace standards.*