Automatic Indoor Localization Algorithm Based on Vision, Geomagnetism and Inertial Navigation

Current domestic and foreign mainstream indoor positioning technology including UWB, Wi-Fi, Bluetooth, RFID they all require intensive infrastructure installation cost. Its difficult to balance in cost, accuracy, stability. Aiming at the above problems, this paper proposes an indoor fusion positioning scheme, which uses image, geomagnetic field and inertial sensor for multi-modal indoor positioning. The method from coarse to fine is adopted to make the full use of the characteristics of each modal signal to locate. Firstly, the base image is selected for image matching to determine the overall range; on this basis, the magnetic field sequence information is filtered by EMD( Empirical Mode Decomposition), and the position is accurately and fine-grained matched by inertial sensor information. This method can be located without the need of additional equipment deployment, and the data information collection is rapid, less affected by the environment, relatively stable, low operation, maintenance costs and good prospects for promotion.