AN AUGMENTED REALITY FRAMEWORK FOR OUTDOOR MOBILE APPLICATIONS

Abstract

This thesis presents the PhD research carried out on mobile Augmented Reality (AR) framework for outdoor applications. Exploiting the sensory instrumentations on the smartphones to achieve the required efficiency and accuracy is the main concern of the sensor-aided descriptor proposed in this thesis. The Efficient Second-order minimization is implemented in a sub-grid manner with an illumination model on mobile device to provide an on-device 3D tracking solution. In addition, image search algorithms are explored and integrated to increase the scalability of the tracking framework. A complete mobile AR framework is delivered with demonstration in real world situations. Search of tracking target is performed on the server by taking advantage of the high computational power of the server workstations, while tracking is performed locally on the mobile device to guarantee real-time 3D tracking and rendering.