README.md

metadata

dataset_info:
  features:
    - name: image
      dtype: image
    - name: crop
      dtype: string
    - name: objects
      struct:
        - name: bbox
          list:
            list: float64
        - name: categories
          list:
            class_label:
              names:
                '0': Blackbean
                '1': Canola
                '2': Corn
                '3': Field Pea
                '4': Flax
                '5': Horseweed
                '6': Kochia
                '7': Lentil
                '8': Ragweed
                '9': Redroot Pigweed
                '10': Soybean
                '11': Sugar beet
                '12': Waterhemp
  splits:
    - name: train
      num_bytes: 7867132445
      num_examples: 1120
  download_size: 7802827920
  dataset_size: 7867132445
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

Weed Crop Detection

A dataset for object detection of weeds and crops in fields. The dataset contains 1,120 images with 17,693 bounding box annotations across 13 categories.

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

Citation

@article{upadhyay2025weed,
  title={Weed-crop dataset in precision agriculture: Resource for AI-based robotic weed control systems},
  author={Upadhyay, Arjun and Mahecha, Maria Villamil and Mettler, Joseph and Howatt, Kirk and Aderholdt, William and Ostlie, Michael and Sun, Xin and others},
  journal={Data in Brief},
  volume={60},
  pages={111486},
  year={2025},
  publisher={Elsevier}
}

Upadhyay, Arjun; G C, Sunil; Villamil Mahecha, Maria; Mettler, Joseph; Howatt, Kirk; Aderholdt, William; Ostlie, Michael; Sun, Xin (2025), “Weed-crop dataset in precision agriculture: Resource for AI-based robotic weed control systems”, Mendeley Data, V2, doi: 10.17632/mthv4ppwyw.2-->