Related Experiment Video
Updated: May 5, 2026
06:21
Micron-scale Phenotyping Techniques of Maize Vascular Bundles Based on X-ray Microcomputed Tomography
Published on: October 9, 2018
8.7K
A cotton organ segmentation method with phenotypic measurements from a point cloud using a transformer.
Fu-Yong Liu1, Hui Geng2, Lin-Yuan Shang2
1College of Information Science and Engineering, Xinjiang University of Science and Technology, Korla, 841000, China.
Plant Methods
|March 17, 2025
Summary
This study introduces TPointNetPlus for segmenting cotton plant 3D point clouds, accurately identifying leaves, bolls, and branches. This method enhances cotton phenomics by enabling precise measurement of key plant traits.
Area of Science:
- Agricultural Science
- Computer Vision
- Plant Biology
Background:
- Accurate measurement of plant phenotypic parameters is essential for cotton growth and development studies.
- 3D point cloud segmentation is a critical step for obtaining precise measurements of plant organ structures.
Purpose of the Study:
- To develop an automated method for cotton organ-level point cloud semantic segmentation.
- To accurately extract phenotypic parameters from segmented cotton plant point clouds.
Main Methods:
- Construction of a dedicated cotton plant point cloud dataset using multi-view images.
- Integration of the Transformer attention module into the PointNet++ model for enhanced feature extraction.
- Application of the HDBSCAN algorithm for organ-level segmentation of leaves, bolls, and branches.
Main Results:
- The TPointNetPlus model achieved 98.39% accuracy in cotton leaf semantic segmentation.
- High correlation coefficients (0.95–0.97) were observed between measured and predicted phenotypic parameters (plant height, leaf area, boll volume).
- Successful segmentation of individual plant organs and extraction of their phenotypic features.
Conclusions:
- The TPointNetPlus method provides accurate and automated analysis of cotton plant 3D point cloud data.
- This approach offers a reliable reference for in-depth plant phenomics research and crop management.
