Indoor Localization Using Camera Phones

Abstract

Indoor localization has long been a goal of pervasive computing research. In this paper, we explore the possibility of determining user's location based on the camera images received from a smart phone. In our system, the smart phone is worn by the user as a pendant and images are periodically captured and transmitted over GPRS to a web server. The web server returns the location of the user by comparing the received images with images stored in a database. We tested our system inside the Computer Science department building. Preliminary results show that user's location can be determined correctly with more than 80% probability of success. As opposed to earlier solutions for indoor localization, this approach does not have any infrastructure requirements. The only cost is that of building an image database.