Computer Science and Engineering, Department of
Date of this Version
2014
Citation
Published in: Proceeding CHI EA '14 CHI '14 Extended Abstracts on Human Factors in Computing Systems, pp. 2173-2178. ACM New York, NY, USA ©2014. ISBN: 978-1-4503-2474-8 doi>10.1145/2559206.2581171.
Abstract
This work in progress describes AerialAR, a global positioning system (GPS) augmented reality (AR) application for mobile devices that automatically labels points of interest (POI) in unmanned aerial vehicle (UAV) imagery. This has important implications for assisting emergency responders. Existing AR applications for UAVs provide the pilot with navigational situational awareness such as terrain features; AerialAR locates and labels mission-relevant points such as schools that may need to be evacuated or hospitals to transport victims to. Locating POI in UAV imagery poses more challenges than those addressed by typical AR browsers on smartphones. The UAV operates at different altitudes as opposed to handheld devices and the UAV camera can tilt over a wide range of angles rather than simply facing forward. AerialAR overcomes these issues by developing a set of equations that translate UAV telemetry and field of view (fov) into a projection onto a Google Map. The map can then be queried for categories of POI. The current version calculates the POI distance and angles with an average error of 0.04% as compared to the Haversine and Rhumb line equations for the distance between the UAV location projected on the ground and the POI on the Google Map. Future work will complete AerialAR by processing UAV video in real-time on mobile devices.
Comments
Copyright 2014. Used by permission.