Deep Learning for Automated Aircraft Surface Defect Detection

Detecting surface defects on aircraft is very critical for ensuring reliability and safety. The complexity of examining aircraft surfaces, coupled with the demand for specialized knowledge, has seen a transformative shift with the advent of deep learning methods. This research provides an in-depth analysis centered on the utilization of Deep Learning (DL) techniques like CNN, YOLO, GAN, and others, in identifying defects on the surfaces of aircraft. Representing the first extensive examination of this emerging field, the review meticulously analyzes twenty-two primary studies, delving into methodological approaches and conceptual foundations. Encompassing major use cases, the study explores their significance in elevating defect detection for aircraft surfaces. Results from this thorough review underscore the potential of deep learning to significantly enhance the accuracy and effectiveness of visual inspection, particularly in identifying defects on aircraft surfaces. Additionally, the study provides a comprehensive compilation of use cases for computer vision and deep neural networks applications in the aviation industry, accompanied by an assessment of technology readiness level. The research brings out the crucial importance of exploring deep learning for detecting aircraft surface defects. These insights drive the aviation industry towards adopting safer and more efficient inspection practices.