Analyzing the Feasibility of PTAM based Localization Algorithms in Large Environments on Mobile Devices

Abstract

Augmented reality (AR) is a young field in science and industry. In AR we interact in a real environment with virtual content. AR has the potential to change how people interact and how they experience their surrounding environment. By combining it with mobile computing, using the power of sophisticated mobile devices, AR will impact almost every industry, particularly retail, entertainment, travel, design, art and advertising. For providing accurate augmentation, precise user localisation is required. This project aims at providing precise localisation of the user in a large indoor environment. We use a real- time tracking and mapping system, which provides an accuracy and robustness comparable to model based methods. The possibility to run tracking and mapping in parallel threads, separately from each other, allows to utilize the power of modern multi-core processors. The resulting system produces detailed maps with thousands of landmarks which can be tracked at interactive frame-rates. The map and camera pose tracking can be used for determining precise user location, for adding augmented content, for tracking a user´s path, for searching for various objects in the environment and for other applications.