add data_card.yaml
Browse files- data_card.yaml +102 -0
data_card.yaml
CHANGED
@@ -0,0 +1,102 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Dataset Metadata
|
2 |
+
|
3 |
+
dataset_info:
|
4 |
+
name: AgriField3D
|
5 |
+
description: >
|
6 |
+
AgriField3D is a curated dataset of 3D point clouds representing fully field-grown maize plants
|
7 |
+
from a diverse maize genetic panel. This dataset contains over 1,000 point clouds of maize plants,
|
8 |
+
collected using a Terrestrial Laser Scanner, and includes various versions of point clouds such as raw,
|
9 |
+
segmented, and reconstructed surfaces. It is designed to support advanced AI applications in agricultural
|
10 |
+
research, particularly maize phenotyping and plant structure analysis.
|
11 |
+
|
12 |
+
version: 1.0
|
13 |
+
license: CC-BY-NC-4.0
|
14 |
+
authors:
|
15 |
+
- Elvis Kimara
|
16 |
+
- Mozhgan Hadadi
|
17 |
+
- Jackson Godbersen
|
18 |
+
- Aditya Balu
|
19 |
+
- Zaki Jubery
|
20 |
+
- Adarsh Krishnamurthy
|
21 |
+
- Patrick Schnable
|
22 |
+
- Baskar Ganapathysubramanian
|
23 |
+
citation: >
|
24 |
+
@article{kimara2025AgriField3D,
|
25 |
+
title = "AgriField3D: A Curated 3D Point Cloud Dataset of Field-Grown Plants from a Maize Diversity Panel",
|
26 |
+
author = "Elvis Kimara, Mozhgan Hadadi, Jackson Godbersen, Aditya Balu, Zaki Jubery, Adarsh Krishnamurthy, Patrick Schnable, Baskar Ganapathysubramanian",
|
27 |
+
year = "2025"
|
28 |
+
}
|
29 |
+
|
30 |
+
intended_use:
|
31 |
+
- AI-based agricultural research
|
32 |
+
- Maize phenotyping
|
33 |
+
- Plant structure analysis
|
34 |
+
- 3D data-driven studies in agriculture
|
35 |
+
|
36 |
+
features:
|
37 |
+
- Point clouds: `.ply` format
|
38 |
+
- Resolutions: 100k, 50k, 10k points
|
39 |
+
- Data types: Raw, segmented, reconstructed surfaces
|
40 |
+
- Plant types: Various maize genetic backgrounds
|
41 |
+
- Segmentation: Individual leaves and stalks color-labeled
|
42 |
+
- Metadata: Quality of point clouds, leaf count, tassels, and maize cobs presence
|
43 |
+
|
44 |
+
dataset_size:
|
45 |
+
raw_point_clouds:
|
46 |
+
- "FielGrwon_ZeaMays_RawPCD_100k.zip: 1045 .ply files (100K points per plant)"
|
47 |
+
- "FielGrwon_ZeaMays_RawPCD_50k.zip: 1045 .ply files (50K points per plant)"
|
48 |
+
- "FielGrwon_ZeaMays_RawPCD_10k.zip: 1045 .ply files (10K points per plant)"
|
49 |
+
segmented_point_clouds:
|
50 |
+
- "FielGrwon_ZeaMays_SegmentedPCD_100k.zip: 520 .ply files (100K points per segmented plant)"
|
51 |
+
- "FielGrwon_ZeaMays_SegmentedPCD_50k.zip: 520 .ply files (50K points per segmented plant)"
|
52 |
+
- "FielGrwon_ZeaMays_SegmentedPCD_10k.zip: 520 .ply files (10K points per segmented plant)"
|
53 |
+
reconstructed_surfaces:
|
54 |
+
- "FielGrwon_ZeaMays_Reconstructed_Surface_stl.zip: 520 .ply files (reconstructed surfaces)"
|
55 |
+
- "FielGrwon_ZeaMays_Reconstructed_Surface_dat.zip: 520 .ply files (NURBS surface data)"
|
56 |
+
|
57 |
+
dependencies:
|
58 |
+
- Python 3.6+
|
59 |
+
- open3d (for visualization)
|
60 |
+
- MeshLab, CloudCompare (for additional point cloud manipulation)
|
61 |
+
- trimesh (for 3D mesh processing)
|
62 |
+
|
63 |
+
installation_instructions: |
|
64 |
+
To install the dataset, clone the repository and install the dependencies:
|
65 |
+
```bash
|
66 |
+
git clone https://huggingface.co/datasets/BGLab/AgriField3D
|
67 |
+
cd AgriField3D
|
68 |
+
pip install -r requirements.txt
|
69 |
+
```
|
70 |
+
|
71 |
+
download_instructions: |
|
72 |
+
1. Download the zipped files from the following links:
|
73 |
+
- FielGrwon_ZeaMays_RawPCD_100k.zip
|
74 |
+
- FielGrwon_ZeaMays_RawPCD_50k.zip
|
75 |
+
- FielGrwon_ZeaMays_RawPCD_10k.zip
|
76 |
+
- FielGrwon_ZeaMays_SegmentedPCD_100k.zip
|
77 |
+
- FielGrwon_ZeaMays_SegmentedPCD_50k.zip
|
78 |
+
- FielGrwon_ZeaMays_SegmentedPCD_10k.zip
|
79 |
+
- FielGrwon_ZeaMays_Reconstructed_Surface_stl.zip
|
80 |
+
- FielGrwon_ZeaMays_Reconstructed_Surface_dat.zip
|
81 |
+
2. Extract the `.zip` files:
|
82 |
+
```bash
|
83 |
+
unzip FielGrwon_ZeaMays_RawPCD_100k.zip
|
84 |
+
unzip FielGrwon_ZeaMays_RawPCD_50k.zip
|
85 |
+
unzip FielGrwon_ZeaMays_RawPCD_10k.zip
|
86 |
+
unzip FielGrwon_ZeaMays_SegmentedPCD_100k.zip
|
87 |
+
unzip FielGrwon_ZeaMays_SegmentedPCD_50k.zip
|
88 |
+
unzip FielGrwon_ZeaMays_SegmentedPCD_10k.zip
|
89 |
+
```
|
90 |
+
|
91 |
+
visualization_instructions: |
|
92 |
+
Use the following Python code to visualize the point clouds:
|
93 |
+
```python
|
94 |
+
import open3d as o3d
|
95 |
+
|
96 |
+
# Load and visualize a PLY file
|
97 |
+
pcd = o3d.io.read_point_cloud("FielGrwon_ZeaMays_RawPCD_100k/0001.ply")
|
98 |
+
o3d.visualization.draw_geometries([pcd])
|
99 |
+
```
|
100 |
+
|
101 |
+
repository_links:
|
102 |
+
- https://huggingface.co/datasets/BGLab/AgriField3D
|