Development of a Low-Cost Computer Vision-Based Omnidirectional Drone Detection System

The increasing proliferation of drones presents significant safety and security challenges, particularly due to their small size and non-metallic materials, which make detection difficult using traditional, high-cost radar systems. This paper proposes a cost-effective, modular, and omnidirectional drone detection system utilizing computer vision techniques. The system, developed with a Raspberry Pi 4B and a multi-camera setup, leverages deep learning algorithms to achieve real-time detection. The proposed system demonstrates a detection precision of 0.53, a confidence score of 0.57, a frame rate of 40 FPS, and an average display delay of 3.70 seconds, capable of detecting drones with a minimum bounding box size of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$40\times 30$</tex> pixels. These results underscore the system's potential for widespread deployment as an affordable, efficient solution for drone detection and perimeter surveillance.