A tracking solution for mobile augmented reality based on sensor-aided marker-less tracking and panoramic mapping

This paper proposes a tracking system for outdoor augmented reality (AR) on handheld devices based on an integration of vision tracking and on-device sensor measurement. To deal with the unpredictable and complex visual information in an outdoor environment, two tracking schemes are proposed for both near-field and far-field tracking scenarios. A sensor-aided binary descriptor is combined with an intensity-based tracking algorithm to deliver a 3D tracking system for fronto-parallel planar surfaces in near-field tracking. In far-field tracking, a sensor-guided panoramic tracking and mapping approach is proposed which allows a creation of the panorama of distant scenes on the fly with camera rotation motion to be tracked at the same time. This implementation allows near real-time creation of panoramic maps on-device; therefore, the users are able to tag information on the training target instantly.