Detection and Localization of Drones and UAVs Using Sound and Vision

Teaser image of drone system

Top left, a schematic of the setup is shown, with a fixed set of nine cameras with a total of 280° field of view, a Pan-Tilt-Zoom camera, and the sound array with twelve microphones. Bottom shows the actual experimental setup. Top right shows two images with a large quadcopter and fixed-wing drone respectively.

Abstract

In this paper, we present a system for drone detection and positioning. The approach uses a system composed of a set of fixed cameras with a large combined field of view (f.o.v.), a Pan-Tilt-Zoom camera and an audio detection rig with a set of microphones. We present how we can define this system in an efficient and modular way, with a number of parallel sensing modules. The focus of the current work is the audio detection and positioning system. We show that we can use direction-of-arrival methods to efficiently position a drone from its ambient emitted sound only, up to several hundred meters distance. The audio system is tested in a real setting, using several different drones, in a number of real flight experiments, with promising results. The dataset with the real audio recordings is available here. Future work includes how to combine the audio processing with the vision system.

Acknowledgement

This work has been supported by Sweden’s Innovation Agency VINNOVA and the Swedish Armed Forces within the programme "Civil-Militärt Innovationsprogram", project nr 2024-03191, the strategic research project ELLIIT and the Wallenberg AI, Autonomous Systems and Software Program (WASP), funded by the Knut and Alice Wallenberg (KAW) Foundation.

BibTeX


@InProceedings{Tegler_2025_CVPR,
  author    = {Tegler, Erik and Modig, Max and Skarin, Per and \r{A}str\"om, Kalle and Oskarsson, Magnus and Flood, Gabrielle},
  title     = {Detection and Localization of Drones and UAVs Using Sound and Vision},
  booktitle = {Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR) Workshops},
  month     = {June},
  year      = {2025},
  pages     = {6650-6658}
}