Visual Communication with UAS: Recognizing Gestures from an Airborne Platform

Current tactical unmanned aerial systems receive their guidance and tasking information predominantly via radio links. To be able to communicate with these systems specific electronic devices are required. This work builds on the concept of visual communication of UAS to allow a person on ground commanding a nearby airborne vehicle to perform a specific reconnaissance task via gestures. A procedure to collect the necessary gestural command components is presented as well as a prototype image processing flow which is able to distinguish between neutral poses, static and dynamic 2D gestures. Prototype experiments prove the applicability of the proposed method on real life data from an airborne platform.