Indoor positioning and obstacle detection for visually impaired navigation system based on LSD-SLAM

Abstract

It is extremely difficult for visually impaired people to walk in strange places, because it is extremely challenging for them to know where they are, the direction in which to move, and the destination. Pedestrian navigation systems such as Google Maps have become popular as a result of the evolution and popularization of smartphones and portable tablet devices. However, it is difficult for visually impaired people to use such systems, because almost all of these are developed for healthy persons. In our research, therefore, we have developed a navigation system for visually impaired people. In order to adapt to dynamic or transient environmental changes, such as temporary obstacles, the system makes use of a small wearable camera to estimate the user's position and construct a 3D environmental map. We applied large-scale direct monocular simultaneous localization and mapping (LSD-SLAM) to develop the system, one of the Visual SLAM techniques, to our system. In this paper, we introduce our approach and system construction, and describe the results of a partial demonstration of its functions.