The growing popularity of unmanned air vehicles (UAVs) for services, such as traffic monitoring, emergency communication, and deliveries, has raised security and privacy concerns due to unauthorized drone use. To address the need for fast, efficient, and precise UAV detection under various conditions, a Lightweight and Computationally Efficient You Only Look Once (LCE-YOLO) architecture is proposed. LCE-YOLO is an enhanced version of YOLOv5s to focus on small and overlooked features critical for robust UAV detection. It is classified to have three variants, each optimized for specific feature maps, reducing computational costs while maintaining accuracy. LCE-YOLO, particularly LCE-YOLO-M, demonstrates significant performance improvements, achieving a precision of 96.8%, recall of 89.2%, mean average precision of 95.9%, and IoU of 50.2% in UAV detection, outperforming state-of-the-art in addressing computational complexity issues.
Lightweight and Computationally Efficient YOLO for Rogue UAV Detection in Complex Backgrounds
IEEE Transactions on Aerospace and Electronic Systems ; 61 , 2 ; 5362-5366
01.04.2025
3424497 byte
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Mob-YOLO: A Lightweight UAV Object Detection Method
IEEE | 2022
|YOLO-SMUG: An Efficient and Lightweight Infrared Object Detection Model for Unmanned Aerial Vehicles
DOAJ | 2025
|