Date of Award
Master of Science in Computer Science
Computer Science and Statistics
Augmented Reality of an outdoor scene as a topic has gained a great deal of popularity in recent years. This work will focus on markerless hybrid Outdoor Augmented Reality (OAR) systems. In general OAR is performed through a classical statistical approach. Strong features are calculated from images of the object, located, and tracked in the scene. Gathering such features requires specialized knowledge of Computer Vision techniques; keeping OAR from finding commercial success. Model-based approaches rely less on previously gathered data, increasingly the accessibility of such techniques, but require extensive scene understanding to correctly parse the scene. The proposed model-based approach minimizes the required scene understanding allowing for augmentation of an environment with minimal input.
Flowers, Brian Adrian, "GRID-BASED OUTDOOR OBJECT RECOGNITION FOR AUGMENTED REALITY" (2014). Open Access Master's Theses. Paper 435.