Related Experiment Video
Updated: Aug 5, 2026
11:34
Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography
Published on: May 15, 2017
Reconstruction-Driven Induction Thermography for AI-Assisted Surface Defect Detection in Welded Structures
Xiang Zhang1,2, Shenghao Huang1, Xiaolu Cui1
1School of Mechatronics and Vehicle Engineering, Chongqing Jiaotong University, Chongqing 400074, China.
Sensors (Basel, Switzerland)
|July 28, 2026
Summary
This study introduces an induction thermography method for inspecting surface defects in steel welds, crucial for preventing fractures in infrastructure. The optimized framework enhances defect detection accuracy for improved structural integrity.
Area of Science:
- Materials Science
- Non-Destructive Testing
- Mechanical Engineering
Background:
- Steel welds are critical in energy storage, wind power, and transportation infrastructure.
- Surface defects in welds can lead to catastrophic failure during service.
- Current inspection methods may not adequately detect critical surface-breaking defects.
Purpose of the Study:
- To develop and validate an induction thermography framework for detecting surface-breaking defects in welded steel structures.
- To optimize excitation structures and reconstruction methods for improved thermal imaging.
- To assess the effectiveness of the framework for automated defect analysis.
Main Methods:
- Optimized a yoke excitation structure using numerical simulation and experimental validation.
- Developed a speed-based reconstruction method to address specimen motion and coil occlusion.
- Evaluated DeepLabv3+, DSCA-UNet, and FPN for automated defect extraction from reconstructed thermograms.
Main Results:
- The optimized yoke configuration provided uniform heating and sufficient intensity in weld regions.
- The speed-based reconstruction method produced continuous temperature fields with clear defect features.
- Automated defect extraction models achieved high precision (up to 92.6%) on the laboratory dataset.
Conclusions:
- The proposed induction thermography framework shows significant potential for post-weld quality inspection.
- The optimized hardware and speed-based reconstruction enhance defect visualization and automated analysis.
- Further validation with larger datasets and diverse weld defects is recommended for real-world application.
