# AgriField3D: A Curated 3D Point Cloud Dataset of Field-Grown Plants From A Maize Diversity Panel

## Dataset Structure

The dataset consists of two compressed `.zip` files, which contain the 3D point cloud data:

```
AgriField3D/
├── FielGrwon_ZeaMays_RawPCD_100k.zip
│   └── Contains 1045 high-resolution (100K points) `.ply` files representing full plant point clouds.
│       ├── 0001.ply
│       ├── 0002.ply
│       ├── ...
│       └── 1045.ply
├── FielGrwon_ZeaMays_RawPCD_50k.zip
│   └── Contains 1045 high-resolution (50K points) `.ply` files representing full plant point clouds.
│       ├── 0001.ply
│       ├── 0002.ply
│       ├── ...
│       └── 1045.ply
├── FielGrwon_ZeaMays_RawPCD_10k.zip
│   └── Contains 1045 high-resolution (10K points) `.ply` files representing full plant point clouds.
│       ├── 0001.ply
│       ├── 0002.ply
│       ├── ...
│       └── 1045.ply
├── FielGrwon_ZeaMays_SegmentedPCD_100k.zip
│   └── Contains 520 high-resolution (100K points) `.ply` files of segmented plant models.
│       ├── 0001.ply
│       ├── 0002.ply
│       ├── ...
│       └── 0520.ply
├── FielGrwon_ZeaMays_SegmentedPCD_50k.zip
│   └── Contains 520 high-resolution (50K points) `.ply` files of segmented plant models.
│       ├── 0001.ply
│       ├── 0002.ply
│       ├── ...
│       └── 0520.ply
├── FielGrwon_ZeaMays_SegmentedPCD_10k.zip
│   └── Contains 520 high-resolution (10K points) `.ply` files of segmented plant models.
│       ├── 0001.ply
│       ├── 0002.ply
│       ├── ...
│       └── 0520.ply
├── Metadata.xlsx
├── PointCloudDownsampler.py
```

### Contents of the `.zip` Files
- **`FielGrwon_ZeaMays_RawPCD_100k.zip`**:
  - Contains 1045 `.ply` files. Each file has 100K point cloud representing an entire maize plant.

- **`FielGrwon_ZeaMays_RawPCD_50k.zip`**:
  - Contains 1045 `.ply` files. Each file has 50K point cloud representing an entire maize plant.

- **`FielGrwon_ZeaMays_RawPCD_10k.zip`**:
  - Contains 1045 `.ply` files. Each file has 10K point cloud representing an entire maize plant.
    
- **`FielGrwon_ZeaMays_SegmentedPCD_100k.zip`**:
  - Contains 520 `.ply` files. Each file represents a segmented maize plant by 100K point cloud focusing on specific plant parts.

- **`FielGrwon_ZeaMays_SegmentedPCD_50k.zip`**:
  - Contains 520 `.ply` files. Each file represents a segmented maize plant by 50K point cloud focusing on specific plant parts.

- **`FielGrwon_ZeaMays_SegmentedPCD_10k.zip`**:
  - Contains 520 `.ply` files. Each file represents a segmented maize plant by 10K point cloud focusing on specific plant parts.
 
### How to Access
1. **Download the `.zip` files**:
   - [FielGrwon_ZeaMays_RawPCD_100k.zip](https://huggingface.co/datasets/BGLab/AgriField3D/resolve/main/FielGrwon_ZeaMays_RawPCD_100k.zip)
   - [FielGrwon_ZeaMays_RawPCD_50k.zip](https://huggingface.co/datasets/BGLab/AgriField3D/resolve/main/FielGrwon_ZeaMays_RawPCD_50k.zip)
   - [FielGrwon_ZeaMays_RawPCD_10k.zip](https://huggingface.co/datasets/BGLab/AgriField3D/resolve/main/FielGrwon_ZeaMays_RawPCD_10k.zip)
   - [FielGrwon_ZeaMays_SegmentedPCD_100k.zip](https://huggingface.co/datasets/BGLab/AgriField3D/resolve/main/FielGrwon_ZeaMays_SegmentedPCD_100k.zip)
   - [FielGrwon_ZeaMays_SegmentedPCD_50k.zip](https://huggingface.co/datasets/BGLab/AgriField3D/resolve/main/FielGrwon_ZeaMays_SegmentedPCD_50k.zip)
   - [FielGrwon_ZeaMays_SegmentedPCD_10k.zip](https://huggingface.co/datasets/BGLab/AgriField3D/resolve/main/FielGrwon_ZeaMays_SegmentedPCD_10k.zip)
2. **Extract the 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
   ```

3. Use the extracted `.ply` files in tools like:
   - MeshLab
   - CloudCompare
   - Python libraries such as `open3d` or `trimesh`.

### Example Code to Visualize the `.ply` Files in Python
```python
import open3d as o3d

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