A famework for detection of drone using YOLOv5x for security surveillance system

Abstract

 In response to the increasing use of drones, there is an emerging need for dependable security surveillance systems capable of detecting them. This article presents a theoretical framework for drone detection using the YOLOv5x (You Only Look Once) deep learning algorithm, which aims to improve security surveillance systems. The framework is composed of various hardware and software components, including drones, cameras, computer systems, and utilizes YOLOv5x to accurately detect drones. To carry out the research, the Roboflow Drone 1 dataset was used, which includes a variety of images captured under various experimental conditions. The YOLOv5x algorithm was selected due to its high accuracy and low computational cost, which makes it an ideal solution for security surveillance systems. To measure the performance of the system, metrics such as F1-Score, Recall, Precision, Area under the Curve, and mAP were employed. Overall, this framework provides a dependable and efficient approach for detecting drones, resulting in improved security of critical infrastructure, public events, and other sensitive locations. Furthermore, the proposed work was compared with existing state-of-the-artwork, and the experimental results demonstrated that the proposed work outperformed the state-of-the-art.