Deep Learning-Based Efficient Drone Detection and Tracking in Real-Time

HOSEN M. I., KAHRAMAN S.

33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025, İstanbul, Türkiye, 25 - 28 Haziran 2025, (Tam Metin Bildiri) identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/siu66497.2025.11111880
  • Basıldığı Şehir: İstanbul
  • Basıldığı Ülke: Türkiye
  • Anahtar Kelimeler: Computer Vision, Deep Learning, DeepSORT, Drone Detection, Drone Tracking, Unmanned Aerial Vehicle, YOLO
  • İstanbul Ticaret Üniversitesi Adresli: Evet

Özet

The increasing deployment of Unmanned Aerial Vehicles (UAVs) across various sectors has raised significant security concerns, as drones can be exploited for unauthorized surveillance, smuggling, and other illicit activities. Addressing these risks necessitates robust and efficient detection and tracking systems that can operate in real-time. While deep learning-based optical detection methods have shown promising results, many existing approaches struggle with low inference speeds, hindering their real-time applicability. In this research, we propose an efficient network for drone detection and tracking that achieves high inference speed. We leverage a lightweight detection model as the base and enhance its performance by incorporating multiple optimization modules. Furthermore, we integrate the DeepSORT algorithm for real-time tracking. Our approach achieves a detection precision of 91.6% and a tracking speed exceeding 40 frames per second, demonstrating its effectiveness for real-time drone surveillance applications.