This study addresses the need for flexibility, reduced human error, and digitalization in Non-Destructive Testing (NDT) within the aerospace sector, focusing on integrating an AI-driven surface defect detection system into the CASPR (Contrôle Automatisé Sur Plateforme Robotique platform in french). CASPR is designed to provide fully automated inspection of large CFRP structures, with AI enhancing defect detection precision. The research presents the development of an AI-based visual detection system for classifying three common surface defects, flakings, scratches, and folds on Airbus components using convolutional neural networks (CNNs). The model was trained on a dataset containing thousands of annotated images of an Airbus part, with some defects artificially generated to improve dataset diversity and ensure comprehensive coverage of the targeted defect types. Data augmentation techniques were employed to prevent overfitting and improve performance on varied data. The model's performance was evaluated using accuracy, precision, recall, and F1-score, with results showing perfect recall (1.0) for all defect types, demonstrating no missed defects. These findings suggest that the AI system significantly enhances quality control efficiency and reliability, reducing dependence on manual inspection and minimizing undetected defects. The successful implementation of this technology indicates potential for widespread adoption in the aerospace manufacturing industry, improving safety and operational efficiency.
NIGET et al. (Fri,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: