A Machine Learning based Automatic defect detection in non stationary thermal wave imaging

The detection of subsurface anomalies is crucial for assessing the structural integrity of materials in various industrial applications. Non-stationary thermal wave imaging has emerged as a reliable non-destructive testing (NDT) technique for identifying defects in composite materials such as glass fiber reinforced polymer (GFRP) and carbon fiber reinforced polymer (CFRP). This research proposes a machine learning-based framework for automatic defect detection using thermal wave imaging data. Supervised learning algorithms, including Decision Tree (DT), Support Vector Machine (SVM), and k-Nearest Neighbor (k-NN), are implemented to classify subsurface defects based on their size and depth. The system is evaluated using signal to-noise ratio (SNR) and detection accuracy metrics to ensure robustness and reliability. The proposed approach enhances defect identification efficiency by leveraging advanced data processing techniques, making it a scalable solution for industrial applications. Experimental results demonstrate improved detection performance compared to conventional defect detection methods