Real-Time Human-Posture Recognition for Human-Drone Interaction Using Monocular Vision

This paper presents a real-time monocular vision solution to human postures recognition for human-drone interaction. The approach achieves a more natural interaction between human and drone. Image regions and joint positions of human bodies in images from a monocular camera mounted on a micro drone are extracted by using a deep neural network. Then, feature vectors of a human body are generated by the relative distance among the joints and classified by a support vector machine (SVM) classifier. The performance of the solution is demonstrated by extensive experiments. Our method obtains an average recognition accuracy of 97.34% on the micro drone and keeps high precision even with a large distance between the human and the micro drone. Furthermore, our method consumes limited computing resources and is suitable for onboard applications of a micro drone.