Related Experiment Video
Updated: Aug 29, 2025
08:25
Using Flatbed Scanners to Collect High-resolution Time-lapsed Images of the Arabidopsis Root Gravitropic Response
Published on: January 25, 2014
12.5K
LiDAR Platform for Acquisition of 3D Plant Phenotyping Database
Manuel G Forero1, Harold F Murcia1, Dehyro Méndez1
1Facultad de Ingeniería, Universidad de Ibagué, Ibagué 730002, Colombia.
Plants (Basel, Switzerland)
|September 9, 2022
Summary
Researchers developed a free database of 3D point clouds and images for maize seedling phenotyping. This platform utilizes 3D Lidar scanning to capture detailed plant morphology, aiding future research.
Area of Science:
- Agricultural Science
- Computer Vision
- Plant Biology
Background:
- Lack of accessible 3D datasets hinders advancements in plant phenotyping.
- 3D imaging offers detailed morphological insights crucial for crop research.
Purpose of the Study:
- To create a publicly available database of 3D point clouds and images for maize seedling phenotyping.
- To establish a standardized platform for high-resolution 3D seedling scanning.
Main Methods:
- Developed a seedling scanning platform using 3D Lidar (SICK LMS4121R-13000) and RGB camera.
- Acquired data from 362 diverse maize seedlings (7-24 days old, 10-40 cm height).
- Processed 3D point clouds for noise reduction (0.03 cm error), synchronized with images, and labeled plant parts.
Main Results:
- Generated a comprehensive database of raw and processed 3D point cloud data with synchronized images.
- Demonstrated database utility with a Random Forest classifier achieving 89.41% accuracy in identifying seedling parts (stem, leaf).
Conclusions:
- The developed platform and database provide a valuable resource for plant phenotyping research.
- The 3D morphological data enables accurate automated analysis of seedling development and structure.
