Datasets:

Project-AgML
/

maize_tomato_weed_classification

dataset_info:
  features:
    - name: image
      dtype: image
    - name: label
      dtype:
        class_label:
          names:
            '0': atriplex
            '1': chenopodium
            '2': convolvulus
            '3': cyperus
            '4': datura
            '5': lolium
            '6': maize
            '7': portulaca
            '8': salsola
            '9': solanum
            '10': sorghum
            '11': tomato
    - name: crop
      dtype: string
    - name: growth_stage
      dtype: string
  splits:
    - name: train
      num_bytes: 95541077
      num_examples: 55828
  download_size: 642109490
  dataset_size: 95541077
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-4.0
task_categories:
  - image-classification
size_categories:
  - 10K<n<100K

Maize Tomato Weed Classification

A dataset for classification of weeds and crops in maize and tomato fields. The dataset contains 55,828 images across 12 classes:
Images per class:

atriplex: 2,459
chenopodium: 3,375
convolvulus: 2,302
cyperus: 3,215
datura: 1,271
lolium: 1,063
maize: 25,273
portulaca: 2,052
salsola: 2,409
solanum: 4,084
sorghum: 1,703
tomato: 6,622

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

Citation

@article{mesias2025drone,
  title={Drone imagery dataset for early-season weed classification in maize and tomato crops},
  author={Mes{\'\i}as-Ruiz, Gustavo A and Pe{\~n}a, Jos{\'e} M and de Castro, Ana I and Dorado, Jos{\'e}},
  journal={Data in Brief},
  volume={58},
  pages={111203},
  year={2025},
  publisher={Elsevier}
}

Mesías-Ruiz, G. A., Peña Barragán, J. M., Castro, A. I. D., & Dorado, J. (2024). DIWEED: Drone Imagery dataset for early-season WEED classification [Data set]. DIGITAL.CSIC. http://doi.org/10.20350/DIGITALCSIC/16559

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