Real-time high-resolution omnidirectional imaging platform for drone detection and tracking
- PDF / 2,286,345 Bytes
- 11 Pages / 595.276 x 790.866 pts Page_size
- 25 Downloads / 183 Views
ORIGINAL RESEARCH PAPER
Real‑time high‑resolution omnidirectional imaging platform for drone detection and tracking Bilal Demir1 · Selman Ergunay1 · Gokcen Nurlu2 · Vladan Popovic1,3 · Beat Ott4 · Peter Wellig4 · Jean‑Philippe Thiran1 · Yusuf Leblebici1 Received: 14 June 2019 / Accepted: 9 October 2019 © Springer-Verlag GmbH Germany, part of Springer Nature 2019
Abstract Drones have become steadily affordable, which raises privacy and security concerns as well as interest in drone detection systems. On the other hand, drone detection is a challenging task due to small dimensions of drones, difficulty of longdistance detection, strict real-time constraints and necessity of wide angle coverage for drones. Although different radar and audio-assisted drone detection systems have been presented, they suffer from the cost, range, or interference problems. On the contrary, a long-range detection can be obtained by a vision-based system. Aiming that, we propose a real-time moving object detection and tracking system optimized for drone detection using 16 cameras with 20 MP resolution. The proposed system detects drones from short range and long range with 360◦ surveillance coverage owing high-performance ultra-highresolution (320 MP) video-processing capability. It is able to detect drones with 100 cm diameter from 700 m distance despite deceptive background. It is interference free, so multiple systems can properly operate in the vicinity without effecting each other. It integrates processing power of embedded systems with flexibility of software to generate a full platform for drone detection and tracking. Keywords Panorama · Background subtraction · High resolution · Moving object detection · Embedded system
1 Introduction Drones are remote-controlled or autonomous unmanned aircraft which are available for public use for different purposes thanks to easy operation. Apart from its benefits, it poses a threat to privacy and security. Hence, drone activities are prohibited in certain areas to ensure safety and to protect privacy. Nevertheless, sole prohibition is not a sufficient solution; therefore, it is necessary to have a detection system to detect unpermitted drones in such areas. Drone market has grown fast, which has provided vast number of small and cheap drones to customers. However, drone detection systems lag behind in this competition, such * Bilal Demir [email protected] 1
École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland
2
Bloomberg LP, London, UK
3
Intel Corp, Santa Clara, CA, USA
4
Armasuisse Science and Technology, Thun, Switzerland
that traditional military radar systems still have limitation of their applicability in general civilian areas [1]. Hence, it has become urgent necessity to build a drone detection system which is capable of overcoming the drone detection challenges such as: • Real-time operation constraint to instantly react to the
upcoming threats.
• Long-distance drone detection to neutralize them before
becoming a serious threat.
• Drone det
Data Loading...
