Related Experiment Video
Updated: Mar 29, 2026
11:34
Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography
Published on: May 15, 2017
11.6K
Enhanced Multi-Scale Defect Detection in Steel Surfaces via Innovative Deep Learning Architecture
1School of Mechanical and Electrical Engineering, Xi'an Technological University, Xi'an 710021, China.
Sensors (Basel, Switzerland)
|March 28, 2026
Summary
This study introduces CTG-YOLO, a deep learning model for steel surface defect detection. It significantly improves accuracy, offering a more reliable solution for industrial quality control.
Area of Science:
- Materials Science
- Computer Science
- Artificial Intelligence
Background:
- Steel surface defects critically affect industrial product quality and safety.
- Existing defect detection methods are often inefficient and limited in scope.
Purpose of the Study:
- To develop an advanced deep learning model for enhanced multi-scale steel surface defect detection.
- To improve the accuracy and efficiency of identifying surface imperfections in steel.
Main Methods:
- An innovative deep learning architecture, CTG-YOLO, was designed.
- The model integrates a CBY parallel network, TFF-PANet neck, and GS-Head detection head.
- Feature extraction and fusion capabilities were enhanced.
Main Results:
- The CTG-YOLO model achieved high mean Average Precision (mAP) scores: 76.55% on NEU-DET and 69.94% on GC10-DET.
- Performance improvements of 3.72% and 3.14% over the original YOLOv8s were recorded.
- Superior feature extraction and fusion were demonstrated.
Conclusions:
- The CTG-YOLO model offers a robust and accurate solution for steel surface defect detection.
- This research lays a strong foundation for practical industrial defect detection applications.
- The developed deep learning approach enhances quality control in steel manufacturing.
