Gesture recognition in the problem of contactless control of an unmanned aerial vehicle

The problem of contactless control of an unmanned aerial vehicle by human gestures is considered. A system of commands is proposed for an AR.Drone 2.0 quadcopter equipped with a builtin computer, two color video cameras, the Linux operating system, and sensors for measuring the flight height, speed, and stability. An Asus Xtion Pro Live three-dimensional sensor based on triangulation and structured light is used to enter data into the control system. Gesture recognition is performed by frame-by-frame processing of a video sequence consisting of depth images. The method does not require initial training, is insensitive to changes in lighting, and is invariant to the sizes of the human palm and body. © 2015, Allerton Press, Inc