Aircraft Defect Localization and Detection Using Object Detection Techniques
This paper presents an aircraft defect detection system based on the YOLO (You Only Look Once) architectures, a state-of-the-art object detection models, and shows that the YOLOv5 and YOLOv8 based approach outperforms existing methods in accuracy and processing speed.
Abstract
Aircraft safety is of paramount importance in the aviation industry, and timely detection of defects or anomalies in aircraft components and structures is crucial for ensuring safe and reliable flight operations. Traditional inspection methods are labor-intensive and prone to human error, making automated defect detection systems a necessity. In this paper, we present an aircraft defect detection system based on the YOLO (You Only Look Once) architectures, a state-of-the-art object detection models. Our proposed framework employs YOLO models starting from version 5 to the recent-one version 8. We detail the data preparation process and present results on various versions of YOLO. Our experiments show that our YOLOv5 and YOLOv8 based approach outperforms existing methods in accuracy and processing speed. We conclude by emphasizing the significance of automated aircraft defect detection in enhancing aviation safety, reducing maintenance costs” and ensuring passenaer well-beina,