AI-Enabled Radar and Camera Integration for Real-Time Drone Detection and Classification

This work introduces a cutting-edge AI framework designed for real-time drone detection, tracking, and classification using radar and camera sensors integrated with edge devices. Through extensive field trials, we demonstrate the system's ability to accurately identify drones and payloads under diverse environmental and operational conditions, both during the day and at night. The dataset includes synchronized radar track, visual, and infrared video, capturing drone maneuvers such as payload drop-offs from up to 1 kilometer away. A key achievement is the successful deployment of our AI algorithms on edge devices, enabling rapid, low-latency processing without the need for cloud computing, making this approach highly relevant for security and infrastructure monitoring applications. Our results show an impressive radar-based drone classification accuracy of 94%, with deep learning models achieving a mean Average Precision (mAP) of 0.58 for daytime detection and 0.61 at night, alongside a payload classification accuracy of 90%.