An Architectural Approach Using Machine Learning for Threat Detection in UAV-Based Defense System

The use of unmanned aerial vehicles (UAVs) has gained popularity during the last 20 years, especially in military applications. UAV use has advanced quickly in the defense industry recently, and several nations now use UAVs for security and surveillance. Object detection is a major obstacle in UAV operations, especially while the UAV is in flight. Real-time detection is made more difficult by variables like object speed and UAV speed. In order to determine if an object is a friend or a foe, this research uses the You Only Look Once (YOLO) deep learning framework for real-time object recognition, sending alerts to the ground station. Moreover, this study presents a comparative analysis among YOLOv5, YOLOv8 and YOLOv11 models, trained with a diverse dataset in various conditions for more precise object detection. This research contributes to advancing drone-based object detection systems and paves the way for future innovations for using UAVs in dynamic and challenging environments.