A Vision based Indoor Navigation system for smartphones in Matlab

We present a new human navigation system on smartphones for indoor environments. The conventional systems have required specific positioning infrastructures installed in buildings. In contrast, the proposed system is not to need such requirements because it can determine the position of a user by image matching to process the images from a smartphone camera indoors like a certain corridor. Although a number of personal positioning methods based on image processing have been proposed, it has been difficult in processing cost to apply them just like that to the proposed system. From this point of view, we have focused on the view-based navigation which has been developed as an indoor positioning method for robots due to its simple algorithm by only one monocular camera. By devising the techniques for applying this view-based navigation to a smartphone, we can visually provide the user having a smartphone the positional and directional information to a destination. This paper describes the algorithm of the view-based navigation speeded up for smartphones and shows experimental results for human navigation performed with usually walking in an indoor corridor.