Update README.md
Browse files
README.md
CHANGED
|
@@ -1,16 +1,14 @@
|
|
| 1 |
-
#
|
| 2 |
|
| 3 |
[**📄 Paper Link**](https://arxiv.org/abs/2503.21958)
|
| 4 |
|
| 5 |
-
---
|
| 6 |
|
| 7 |
## 🧾 Overview
|
| 8 |
|
| 9 |
This dataset complements our paper on a stationary-camera-based NeRF framework for high-throughput 3D plant phenotyping. Traditional NeRF pipelines require a moving camera around a static object, which is impractical in automated indoor phenotyping environments. Our method enables 3D point cloud (PCD) reconstruction using a stationary camera and a rotating object, significantly simplifying the imaging setup.
|
| 10 |
|
| 11 |
-
The **
|
| 12 |
|
| 13 |
-
---
|
| 14 |
|
| 15 |
## 📁 Dataset Directory Structure
|
| 16 |
AgriPCD/
|
|
@@ -39,9 +37,8 @@ Includes COLMAP pose estimations and sparse point clouds generated using the `ns
|
|
| 39 |
Stores trained NeRF models for each object using the `nerfacto` trainer in Nerfstudio.
|
| 40 |
|
| 41 |
### `pcd/`
|
| 42 |
-
Final 10M-point cloud representations of each object
|
| 43 |
|
| 44 |
-
---
|
| 45 |
|
| 46 |
## 🌿 Object Categories
|
| 47 |
|
|
@@ -59,8 +56,6 @@ Each object is reconstructed under two experimental protocols:
|
|
| 59 |
- `SC` – Stationary Camera, Rotating Object (Primary Contribution)
|
| 60 |
- `GT` – Ground Truth: Moving Camera, Stationary Object (Reference)
|
| 61 |
|
| 62 |
-
---
|
| 63 |
-
|
| 64 |
## 🎥 Imaging Protocol
|
| 65 |
|
| 66 |
- Captured using an **iPhone 13 Mini** at 4K, 30fps
|
|
@@ -68,8 +63,6 @@ Each object is reconstructed under two experimental protocols:
|
|
| 68 |
- **ArUco markers** and a **ping pong ball (Ø = 0.04 m)** used for pose and scale calibration
|
| 69 |
- Processed using **COLMAP** and **Nerfstudio**
|
| 70 |
|
| 71 |
-
---
|
| 72 |
-
|
| 73 |
## 🧪 Applications
|
| 74 |
|
| 75 |
- High-throughput plant phenotyping
|
|
@@ -77,14 +70,11 @@ Each object is reconstructed under two experimental protocols:
|
|
| 77 |
- Point cloud alignment and evaluation
|
| 78 |
- Hyperspectral and multimodal NeRF fusion
|
| 79 |
|
| 80 |
-
---
|
| 81 |
-
|
| 82 |
## 📜 License
|
| 83 |
|
| 84 |
**CC-BY-NC-4.0**
|
| 85 |
[Creative Commons Attribution-NonCommercial 4.0 International License](https://creativecommons.org/licenses/by-nc/4.0/)
|
| 86 |
|
| 87 |
-
---
|
| 88 |
## 🔖 Citation
|
| 89 |
If you use this dataset in your work, please cite:
|
| 90 |
|
|
|
|
| 1 |
+
# SC-NeRF: A Dataset for Stationary-Camera-Based 3D Reconstruction Using Neural Radiance Fields
|
| 2 |
|
| 3 |
[**📄 Paper Link**](https://arxiv.org/abs/2503.21958)
|
| 4 |
|
|
|
|
| 5 |
|
| 6 |
## 🧾 Overview
|
| 7 |
|
| 8 |
This dataset complements our paper on a stationary-camera-based NeRF framework for high-throughput 3D plant phenotyping. Traditional NeRF pipelines require a moving camera around a static object, which is impractical in automated indoor phenotyping environments. Our method enables 3D point cloud (PCD) reconstruction using a stationary camera and a rotating object, significantly simplifying the imaging setup.
|
| 9 |
|
| 10 |
+
The **SC-NeRF** dataset includes videos, extracted frames, COLMAP pose estimates, NeRF training outputs, and final 3D point clouds. It supports the community in replicating, extending, or evaluating methods for indoor phenotyping under controlled imaging conditions with minimal hardware requirements.
|
| 11 |
|
|
|
|
| 12 |
|
| 13 |
## 📁 Dataset Directory Structure
|
| 14 |
AgriPCD/
|
|
|
|
| 37 |
Stores trained NeRF models for each object using the `nerfacto` trainer in Nerfstudio.
|
| 38 |
|
| 39 |
### `pcd/`
|
| 40 |
+
Final 10M-point cloud representations of each object were filtered, scaled using metric references, and aligned with ground-truth models for evaluation.
|
| 41 |
|
|
|
|
| 42 |
|
| 43 |
## 🌿 Object Categories
|
| 44 |
|
|
|
|
| 56 |
- `SC` – Stationary Camera, Rotating Object (Primary Contribution)
|
| 57 |
- `GT` – Ground Truth: Moving Camera, Stationary Object (Reference)
|
| 58 |
|
|
|
|
|
|
|
| 59 |
## 🎥 Imaging Protocol
|
| 60 |
|
| 61 |
- Captured using an **iPhone 13 Mini** at 4K, 30fps
|
|
|
|
| 63 |
- **ArUco markers** and a **ping pong ball (Ø = 0.04 m)** used for pose and scale calibration
|
| 64 |
- Processed using **COLMAP** and **Nerfstudio**
|
| 65 |
|
|
|
|
|
|
|
| 66 |
## 🧪 Applications
|
| 67 |
|
| 68 |
- High-throughput plant phenotyping
|
|
|
|
| 70 |
- Point cloud alignment and evaluation
|
| 71 |
- Hyperspectral and multimodal NeRF fusion
|
| 72 |
|
|
|
|
|
|
|
| 73 |
## 📜 License
|
| 74 |
|
| 75 |
**CC-BY-NC-4.0**
|
| 76 |
[Creative Commons Attribution-NonCommercial 4.0 International License](https://creativecommons.org/licenses/by-nc/4.0/)
|
| 77 |
|
|
|
|
| 78 |
## 🔖 Citation
|
| 79 |
If you use this dataset in your work, please cite:
|
| 80 |
|