Related Experiment Video
Updated: Apr 13, 2026
11:49
Using High Resolution Computed Tomography to Visualize the Three Dimensional Structure and Function of Plant Vasculature
Published on: April 5, 2013
21.0K
A 3D reconstruction platform for complex plants using OB-NeRF
Sixiao Wu1, Changhao Hu1, Boyuan Tian1
1China College of Engineering/Key Laboratory of Agricultural Equipment for the Middle and Lower Reaches of the Yangtze River, Ministry of Agriculture, Huazhong Agricultural University, Wuhan, China.
Frontiers in Plant Science
|March 25, 2025
Summary
This study introduces a fast and accurate 3D plant reconstruction platform using Object-Based NeRF (OB-NeRF). The system significantly reduces reconstruction time and enhances precision for agricultural digital twins and high-throughput phenotyping.
Area of Science:
- Agricultural Science
- Computer Vision
- 3D Reconstruction
Background:
- 3D reconstruction enhances high-throughput phenotyping and digital twin development in agriculture.
- Existing 3D reconstruction platforms face challenges with cost, speed, workflow complexity, and automation.
Purpose of the Study:
- To develop an efficient and automated 3D reconstruction platform for complex plants.
- To overcome limitations of existing 3D reconstruction methods in terms of speed, cost, and automation.
Main Methods:
- A video acquisition system and keyframe extraction were implemented.
- Zhang Zhengyou's calibration and Structure from Motion (SfM) were used for camera pose estimation.
- Object-Based NeRF (OB-NeRF) with a novel ray sampling strategy was employed for fine-scale plant reconstruction.
Main Results:
- The OB-NeRF platform achieved high-quality 3D plant reconstruction in 250 seconds.
- OB-NeRF demonstrated superior performance in PSNR evaluation and reduced reconstruction time compared to NeRF and other methods.
- The reconstructed 3D models showed high fidelity, with R² values of 0.9933, 0.9881, and 0.9883 for plant height, leaf length, and leaf width, respectively.
Conclusions:
- The developed 3D reconstruction platform offers a robust solution for agricultural digital twins and high-throughput phenotyping.
- The OB-NeRF algorithm significantly improves reconstruction efficiency and quality without background segmentation.
- Optimized camera poses and automated localization enhance the precision and applicability of the 3D models.
