Abstract
Indoor position technology based on visual becames a hot spot which can provide high position accuracy and more scene information. The system proposed by this paper is operated in real time, robust to severe motion clutter and brightness transformation and low cost which combines several steps include image acquisition and correction, object recognition based on deep learning and stereo matching. And this paper also proposes a new multi-feature fusion feature point extraction and matching algorithm for object accurate location. Finally, we present an evaluation in the dynamic indoor positioning tests and the result of the indoor horizontal positioning error is less than 1m.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
