Related Experiment Video
Updated: Jun 15, 2025
11:49
Cereal Crop Ear Counting in Field Conditions Using Zenithal RGB Images
Published on: February 2, 2019
9.3K
Efficient online detection device and method for cottonseed breakage based on Light-YOLO.
Hongzhou Zhang1, Qingxu Li2, Zhenwei Luo1
1College of Mechanical and Electrical Engineering, Tarim University, Alar, China.
Frontiers in Plant Science
|August 26, 2024
Summary
This study introduces a new system for detecting broken cottonseed, crucial for successful cotton planting. The developed Light-YOLO model offers a fast, cost-effective solution for identifying damaged seeds, improving crop yields.
Area of Science:
- Agricultural Engineering
- Computer Vision
- Machine Learning
Background:
- High-quality cottonseed is vital for optimal cotton production and germination.
- Current methods for detecting broken cottonseed lack speed and cost-effectiveness.
- Seed integrity directly impacts plant growth and yield.
Purpose of the Study:
- To develop a rapid and cost-effective online detection system for broken cottonseed.
- To improve the accuracy and efficiency of cottonseed breakage detection using deep learning.
- To validate the performance of the developed system in real-world conditions.
Main Methods:
- A dual-camera system was designed for capturing front and back images of cottonseeds.
- Hardware, software, and control systems were developed for online detection.
- The YOLOv8m model was enhanced with MobileOne-block and GhostConv, creating Light-YOLO.
- Light-YOLO was trained and evaluated for cottonseed breakage detection.
Main Results:
- Light-YOLO achieved 93.8% precision, 97.2% recall, 98.9% mAP50, and 96.1% accuracy with a 41.3 MB model size.
- YOLOv8m achieved 93.7% precision, 95.0% recall, 99.0% mAP50, and 95.2% accuracy with a 49.6 MB model size.
- Online validation demonstrated an 86.7% detection accuracy for cottonseed breakage.
Conclusions:
- Light-YOLO offers superior detection performance and speed compared to YOLOv8m for cottonseed breakage.
- The developed online detection technology is feasible and effective for sorting broken cottonseeds.
- This innovation provides a practical solution for improving cottonseed quality in agriculture.
