An Indoor Positioning Method Based on Smartphone and Surveillance Camera
Ao Liu, Wenguang Wang
Abstract
Indoor pedestrian positioning using the surveillance camera may suffer from limited coverage. A novel method based on the smartphone and surveillance camera is proposed to locate accurately and continuously when a pedestrian moves in an area with limited camera coverage. In the area covered by the camera, to reduce the long-distance error, a D-S evidence theory-based weighted fusion (DS-WF) method is proposed, which can fuse passive visual positioning and pedestrian dead reckoning (PDR). In the blind area, to solve the problem of geomagnetic mismatching, a two-stage geomagnetic positioning (TS-GP) method that utilizes the geomagnetic intensity and geomagnetic features together is proposed. Then, the extended Kalman filter (EKF) is used to fuse the positioning results from TS-GP and PDR. The experiments are conducted in the scenery including four corridors and three turns to evaluate the pedestrian positioning performance of the proposed method. The experimental results show that the proposed method can achieve stable and continuous positioning when a pedestrian pass through different areas, namely, the camera coverage area or blind area.