# Dataset Metadata

dataset_info:
  name: AgriField3D
  description: >
    AgriField3D is a curated dataset of 3D point clouds representing fully field-grown maize plants 
    from a diverse maize genetic panel. This dataset contains over 1,000 point clouds of maize plants, 
    collected using a Terrestrial Laser Scanner, and includes various versions of point clouds such as raw, 
    segmented, and reconstructed surfaces. It is designed to support advanced AI applications in agricultural 
    research, particularly maize phenotyping and plant structure analysis.

  version: 1.0
  license: CC-BY-NC-4.0
  authors:
    - Elvis Kimara
    - Mozhgan Hadadi
    - Jackson Godbersen
    - Aditya Balu
    - Zaki Jubery
    - Adarsh Krishnamurthy
    - Patrick Schnable
    - Baskar Ganapathysubramanian
  citation: >
    @article{kimara2025AgriField3D,
        title = "AgriField3D: A Curated 3D Point Cloud Dataset of Field-Grown Plants from a Maize Diversity Panel",
        author = "Elvis Kimara, Mozhgan Hadadi, Jackson Godbersen, Aditya Balu, Zaki Jubery, Adarsh Krishnamurthy, Patrick Schnable, Baskar Ganapathysubramanian",
        year = "2025"
    }

  intended_use:
    - AI-based agricultural research
    - Maize phenotyping
    - Plant structure analysis
    - 3D data-driven studies in agriculture

  features:
    - Point clouds: `.ply` format
    - Resolutions: 100k, 50k, 10k points
    - Data types: Raw, segmented, reconstructed surfaces
    - Plant types: Various maize genetic backgrounds
    - Segmentation: Individual leaves and stalks color-labeled
    - Metadata: Quality of point clouds, leaf count, tassels, and maize cobs presence

  dataset_size: 
    raw_point_clouds: 
      - "FielGrwon_ZeaMays_RawPCD_100k.zip: 1045 .ply files (100K points per plant)"
      - "FielGrwon_ZeaMays_RawPCD_50k.zip: 1045 .ply files (50K points per plant)"
      - "FielGrwon_ZeaMays_RawPCD_10k.zip: 1045 .ply files (10K points per plant)"
    segmented_point_clouds: 
      - "FielGrwon_ZeaMays_SegmentedPCD_100k.zip: 520 .ply files (100K points per segmented plant)"
      - "FielGrwon_ZeaMays_SegmentedPCD_50k.zip: 520 .ply files (50K points per segmented plant)"
      - "FielGrwon_ZeaMays_SegmentedPCD_10k.zip: 520 .ply files (10K points per segmented plant)"
    reconstructed_surfaces:
      - "FielGrwon_ZeaMays_Reconstructed_Surface_stl.zip: 520 .ply files (reconstructed surfaces)"
      - "FielGrwon_ZeaMays_Reconstructed_Surface_dat.zip: 520 .ply files (NURBS surface data)"

  dependencies:
    - Python 3.6+
    - open3d (for visualization)
    - MeshLab, CloudCompare (for additional point cloud manipulation)
    - trimesh (for 3D mesh processing)

  installation_instructions: |
    To install the dataset, clone the repository and install the dependencies:
    ```bash
    git clone https://huggingface.co/datasets/BGLab/AgriField3D
    cd AgriField3D
    pip install -r requirements.txt
    ```

  download_instructions: |
    1. Download the zipped files from the following links:
       - FielGrwon_ZeaMays_RawPCD_100k.zip
       - FielGrwon_ZeaMays_RawPCD_50k.zip
       - FielGrwon_ZeaMays_RawPCD_10k.zip
       - FielGrwon_ZeaMays_SegmentedPCD_100k.zip
       - FielGrwon_ZeaMays_SegmentedPCD_50k.zip
       - FielGrwon_ZeaMays_SegmentedPCD_10k.zip
       - FielGrwon_ZeaMays_Reconstructed_Surface_stl.zip
       - FielGrwon_ZeaMays_Reconstructed_Surface_dat.zip
    2. Extract the `.zip` files:
       ```bash
       unzip FielGrwon_ZeaMays_RawPCD_100k.zip
       unzip FielGrwon_ZeaMays_RawPCD_50k.zip
       unzip FielGrwon_ZeaMays_RawPCD_10k.zip
       unzip FielGrwon_ZeaMays_SegmentedPCD_100k.zip
       unzip FielGrwon_ZeaMays_SegmentedPCD_50k.zip
       unzip FielGrwon_ZeaMays_SegmentedPCD_10k.zip
       ```

  visualization_instructions: |
    Use the following Python code to visualize the point clouds:
    ```python
    import open3d as o3d

    # Load and visualize a PLY file
    pcd = o3d.io.read_point_cloud("FielGrwon_ZeaMays_RawPCD_100k/0001.ply")
    o3d.visualization.draw_geometries([pcd])
    ```

  repository_links:
    - https://huggingface.co/datasets/BGLab/AgriField3D
    - https://huggingface.co/docs/hub/datasets-cards