Related Experiment Video
Updated: May 29, 2025
11:34
Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography
Published on: May 15, 2017
11.1K
Surface defect detection on industrial drum rollers: Using enhanced YOLOv8n and structured light for accurate
Guofeng Qin1,2, Qinkai Zou2, Mengyan Li2
1Teachers College for Vocational and Technical Education, Guangxi Normal University, Guilin, People's Republic of China.
Plos One
|February 5, 2025
Summary
This study introduces a new line fringe structured light system and an enhanced YOLOv8n deep learning model for detecting drum roller surface defects. The improved system achieves high accuracy (97.2% mAP) and speed (4.3ms), ensuring better production quality.
Area of Science:
- Industrial Automation
- Computer Vision
- Machine Learning
Background:
- Traditional visual inspection systems struggle with drum roller surface defect detection due to unclear feature reflection, low efficiency, and insufficient accuracy.
- Existing methods lack the precision required for effective quality control in drum roller manufacturing.
Purpose of the Study:
- To design an advanced image acquisition system using line fringe structured light for clearer defect visualization.
- To develop an improved deep learning model based on YOLOv8n for efficient and accurate drum roller surface defect detection.
Main Methods:
- Implemented a line fringe structured light imaging system to enhance defect feature visibility.
- Adapted the YOLOv8n model with deformable convolution, a novel feature fusion module, and Wise-IoU loss for superior performance.
- Utilized deformable convolution to improve feature extraction and a new fusion module for richer information integration.
Main Results:
- The enhanced YOLOv8n model achieved a mean average precision (mAP) of 97.2% for defect detection.
- The system demonstrated a rapid detection time of 4.3ms, indicating high efficiency.
- The proposed system and algorithm significantly improved the accuracy and efficiency of drum roller surface defect detection.
Conclusions:
- The developed line fringe structured light imaging system and improved YOLOv8n model effectively enhance drum roller production quality.
- The system offers a robust solution for real-time, accurate surface defect detection in industrial settings.
- Future research should consider rotated bounding boxes and diverse datasets to address limitations with elongated defects and improve generalization.
