A cross-platform smart drone controller framework - for real-time surveillance and inspection

Abstract

This can form the basis for strategic management of new Smart Cities that aim to reduce Line-of-Sight piloting overheads and automate surveillance task. In this paper, we propose an architectural design to support the needed capabilities and basic features of a Smart Drone Controller. framework. This prototype framework supports a deployed team of Wi-Fi drones to conduct assigned surveillance. SDC's machine learning engine evaluated with Deep Learning algorithms to detect target accuracy from drones live video feed. We have used Darknet/YOLO model for custom object detection in Intruder drone and Wall crack detection use cases and accuracy is reaching over 80%, which is encouraging.