Title: Tracking for augmented reality applications on mobile phones
Thesis by: Frieder Pankratz
Advisor: Prof. Gudrun Klinker, PhD.
Supervision by: Björn Schwerdtfeger
, Peter Keitler
, Daniel Pustka
Due date: 15.07.2009
Mobile phones offer more and more gaphic and computing power. They are widespread and most people are comfortable using them. That's why mobile phones are getting more interesting as a platform for Augmented Reality(AR) applications. Even through they offer only a fraction of the processing power of a desktop pc, they offer sufficient processing power for different tracking techniques. The tracking techniques used for desktop computing can't simply be applied to mobile phones. The used techniques must be, among other things, computing and memory efficient. In this master's thesis different techniques for detecting and tracking objects or camera movement in all six degrees of freedom for AR applications are presented. They are evaluated for their use on mobile phones. Among other technical restrictions, the user requirements for the AR system had to be considered (robustness, speed). Small user studies have been performed to observe the behavior and needs of the users, while using a mobile AR application. In respect to these requirements, a marker tracking system was extended, to allow the loss of the marker. If the marker gets lost, the movements of the mobile phone are approximated, using the horizontal and vertical movement of the camera image. The new marker pose is estimated using the movements of the mobile phone and the last known marker pose. Thereby the user should gain more mobility, which should allow him to interact with the AR application more naturally, compared to a standard marker tracking system.