Deep Learning-Based Defect Detection for Light Aircraft with Unmanned Aircraft Systems (UAS)

Abstract

Aircraft safety and maintenance have long relied on manual inspections, which are time-consuming and prone to human error. This paper presents a deep learning-based automated defect detection system that uses Unmanned Aircraft Systems (UAS) to inspect light aircraft. High-resolution images captured by UAS are analyzed using Convolutional Neural Networks (CNNs) to detect defects such as cracks, corrosion, and surface deformation. This system offers real-time, consistent, and non-invasive inspection capabilities. Through the integration of image acquisition technology and AI-driven analysis, the proposed approach significantly reduces inspection time, enhances detection accuracy, and improves safety in aviation maintenance operations. A comparative evaluation with traditional inspection methods shows the system’s effectiveness in identifying defects with over 90% accuracy across various aircraft surface types. The results affirm that deep learning, combined with drone technology, can provide a reliable, efficient, and scalable solution for aircraft defect detection.