Design and Development of Drone Detection System using YOLOv11

Abstract

The rapid growth of Unmanned Aerial Vehicles (UAVs) has created significant opportunities across applications such as surveillance, agriculture, logistics, and infrastructure monitoring. However, the increasing use of drones has also introduced security and privacy concerns, particularly in sensitive and restricted areas. To address these challenges, this paper presents a real-time drone detection and classification framework based on the YOLOv11 deep learning model. A custom dataset containing three drone categories, namely Fixed-wing, Multicopter, and Vertical Take-Off and Landing (VTOL), was developed and annotated for training and evaluation. Image preprocessing and data augmentation techniques, including resizing, rotation, scaling, brightness adjustment, and horizontal flipping, were applied to improve model robustness under diverse environmental conditions. The YOLOv11 model was trained to detect drones and generate bounding boxes, class labels, and confidence scores. Furthermore, a classification logic was incorporated to determine whether a detected drone belongs to the trained dataset categories. Experimental results show that the model achieves reliable detection accuracy. The proposed framework provides an efficient, automated, and scalable solution for drone surveillance and security applications, while maintaining a balance between detection accuracy and computational efficiency.