Urban Positioning Using Smartphone-Based Imaging

Orientation and navigation in a world dominated by visual signs is still a major problem for blind and visually impaired people. The Global Positioning System is of limited use due to its inaccuracy particularly in urban environments. Therefore, we propose a novel approach of precise localization on predefined routes with the help of smartphones and image processing techniques. From an initial acquisition of a given route, a three-dimensional reconstruction is created. A query image is submitted to the database and the location and direction of the camera are calculated. Here, we demonstrate our approach on a evaluationdataset with a mean positioning error of 5.51meters.