Image Based Navigation System for Pedestrians in an Indoor Environment

Abstract

Indoor navigation systems provide means to guide pedestrians to their various destinations. While many tools that take advantage of the Global Positioning System (GPS) for outdoor navigation exists, their usefulness is limited to the availability of GPS signal reception, which is usually poor in indoor environments. In this research, we propose a method that employs only images for indoor pedestrian navigation. In the proposed method, a map of the indoor environment is first transformed into a graph model where features of indoor environment are attached to graph nodes and their distances represented by the graph edges. Feature images of initial and destination locations are provided by the pedestrian who needs to be guided. These images are fed into the created graph model and thereafter the Speeded-Up Robust Features (SURF) is then used to find a match to these images to discover their corresponding graph nodes. Graph nodes are identified in a manner that corresponds to pedestrian localized position and destination. Leveraging on these nodes in the graph, models are proposed to find the shortest path to user’s destination with instructions and graphical navigation path to enhance maneuverability. Experiment carried out on an indoor environment of the French South African Institute of Technology building, (Tshwane University of Technology) shows encouraging results.