Computer vision and sensor fusion for efficient hybrid tracking in augmented reality systems

Abstract

Augmented Reality is one of the most fundamental enabling technologies of this era which modifies our perception to such an extent that we are able to see, hear and feel the ordinary everyday objects in a new and enriched way. This technology is an amalgamation of Real and Virtual Worlds which aims to enhance the physical information by exactly super-imposing the computer generated content on it. The technology can prove helpful to surgeons and doctors to visualize and train a surgery, can instruct a mechanic for repair of an unknown piece of equipment, can serve as a source of entertainment, can help soldiers for spotting enemy snipers, and can be used as an efficient teaching aid for education. A typical Augmented Reality (AR) system requires the integration of hardware and software with factors such as Tracking, Registration, Interaction and Display playing an important role in the system design. The focus of this paper is specifically towards surveying the available Tracking Techniques for various Augmented Reality applications. Based on the existing methods of sensor based, vision based and hybrid tracking, a new method for efficient tracking is proposed.