Research On Key Techniques Of Tracking And Registration For Mobile Augmented Reality

Augmented Reality is a very promising technology, which can realize the seamlessintegration between the virtual information and the real world, and provide us a newenvironment with real sensory effects. Along with the rapid development of mobile terminals,their ability of computing, multimedia and network accessing have made a difference. Besides,its built-in camera, GPS and electronic compass and other sensors have provided a strongguarantee for precise positioning and Augmented Reality.As the key technology of Augmented Reality, registration and tracking must not onlyguarantee a certain accuracy and stability, but also be able to reduce the amount of calculationand improve the operating speed. In this thesis, we focus on the tracking and registrationtechnology of Augmented Reality, and then propose some appropriate solutions finally.First of all, the camera model and calibration technology are analyzed in the article. Thenthe coordinate transformation and registration between virtual information and reality sceneare discussed, and several classical algorithms for natural feature recognition are analyzed.Secondly, a fast algorithms for natural feature recognition: SUORB, are proposed, whichcombines an E-FAST feature detection algorithm and a dynamic length ORB descriptor. In thefeature point detection phase, an enhanced algorithm based on the FAST feature detectionalgorithm: E-FAST is proposed. This algorithm can improve the detection speed and keep therobustness. Meanwhile, a dynamic length ORB descriptor supporting the E-FAST algorithm isproposed, which can reduce computational complexity and memory usage when describingkeypoints, and improve matching efficiency. The experimental results show that the E-FASTalgorithm can get high performances such as a small amount of computation, high efficiency,and good robustness. It can be well applied to mobile Augmented Reality.Finally, an urban shopping guide application based on hybrid tracking technology on theAndroid smart phone are implemented. This application use GPS or network parameter forlocation and electronic compass for orientation, and use the SUORB algorithm to distinguishand track business logos, then display related information on the screen, and complete theintegration of the virtual information and the real world.