A Machine Learning Based Method for Object Detection and Localization Using a Monocular RGB Camera Equipped Drone

Growing Unmanned Aerial Vehicle (UAV) market trends and interest in potential uses such as monitoring, visual inspection, object detection, and path planning have shown promising results using machine learning techniques. However, UAV adoption faces several challenges in real-life scenarios as low-accuracy sensors are involved in the identification, tracking, and localization of UAVs. In order to overcome the aforementioned challenges, this paper proposes an intelligent machine learning-based system coupled with computer vision (CV) to detect objects and localize UAVs equipped with just a monocular camera. The experimental results using the Telo DJI drone demonstrate that the proposed methodology can detect, track objects, and localize the drone with high accuracy. The system's ability for automated monitoring in real environments can lend its uses for urban traffic, logistics, and security applications.