Lucid: Lidar-Based Uav Classification and Detection for Airspace Surveillance

Citations

WEB OF SCIENCE

0
Citations

SCOPUS

0

초록

Misuse of unmanned aerial vehicle (UAV) has surged dramatically in recent years, particularly in warfare and terrorism. These UAVs are small, and often operate at nights, making them challenging to detect solely using radar or vision-based surveillance systems. To address this concern, we propose Lucid, a LiDAR-based UAV detection system designed to identify fastmoving UAVs even in complex environments. Lucid distinguish UAVs from other objects by analyzing their unique mobility characteristics using spatio-temporal background subtraction, flying object extraction, flight trajectory estimation, and point cloud fusion techniques. We conduct extensive real-world LiDAR measurements using three different types and sizes of UAVs at four distinct environments to obtain 3D point cloud datasets of flight trajectories, and evaluate Lucid to demonstrate that it can reliably detect and distinguish UAVs from other flying objects such as birds with 96.39% accuracy. Lucid offers a practical and effective approach for detection of UAV threats, and can be used together with existing surveillance systems to address the growing concerns over UAV misuse.

키워드

3D Point Cloud; Classification; Drone Detection; LiDAR; Surveillance; Unmanned Aerial Vehicle; UAV
제목
Lucid: Lidar-Based Uav Classification and Detection for Airspace Surveillance
저자
Kim, Deokjin; Paek, Jeongyeup; Bahk, Saewoong
DOI
10.1109/2575-8411.2026.00095
발행일
2026
유형
Proceedings Paper
저널명
Proceedings - International Conference on Distributed Computing Systems
페이지
949 ~ 959