---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: mask
    dtype: image
  - name: split
    dtype: string
  - name: date
    dtype: string
  - name: time
    dtype: string
  - name: lat
    dtype: string
  - name: lon
    dtype: string
  - name: utc
    dtype: string
  - name: weather
    dtype: string
  - name: wind
    dtype: string
  - name: seeding_date
    dtype: string
  - name: height_mm
    dtype: string
  - name: depth
    dtype: image
  - name: depth_refined
    dtype: string
  splits:
  - name: train
    num_bytes: 9013751981
    num_examples: 2568
  download_size: 8747620007
  dataset_size: 9013751981
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: cc-by-4.0
task_categories:
- image-segmentation
size_categories:
- 1K<n<10K
---
# We3ds Weed Segmentation

This dataset provides real-world RGB imagery focused on weed segmentation within agricultural fields. It captures diverse weed and crop patterns under natural field conditions, offering a practical resource for developing and evaluating semantic segmentation models in precision agriculture applications. The dataset contains 2,568 images with pixel-level mask annotations.

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

The original train/test/val split has been preserved in the `split` column.

## Citation

```bibtex
@article{kitzler2023we3ds,
  title={WE3DS: An RGB-D image dataset for semantic segmentation in agriculture},
  author={Kitzler, Florian and Barta, Norbert and Neugschwandtner, Reinhard W and Gronauer, Andreas and Motsch, Viktoria},
  journal={Sensors},
  volume={23},
  number={5},
  pages={2713},
  year={2023},
  publisher={MDPI}
}
```

Kitzler, F., Barta, N., Neugschwandtner, R. W., Gronauer, A., & Motsch, V. (2023). WE3DS: An RGB-D image dataset for semantic segmentation in agriculture [Dataset]. In Sensors (Vol. 23, Issue 5, p. 2713). Zenodo. https://doi.org/10.5281/zenodo.7457983

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