Indoor Position Algorithm Based on the Fusion of Wifi and Image

Abstract

In this paper, an image and WiFi multi-mode fusion localization method is proposed. WiFi fingerprint 10-calization is one of the most commonly used methods at present, but WiFi fingerprint localization has the phenomena of multipath and non-line-of-sight, signal fluctuation, scattering and so on, which leads to the unstable positioning effect. In addition, the image uses feature matching to determine the user’s location, which is less affected by the environment, more stable and lower cost. Taking advantage of the advantages of the two algorithms, a new fusion location estimation algorithm is proposed. In the offline phase, WiFi fingerprint and image are collected, fingerprint database is established, and the collected images are split and matched using AlexNet to determine the overall range. In the online stage, fingerprint location matching is carried out in the area through the fusion WiFi. After experiments, the results show that our method can effectively reduce the fluctuation of WiFi fingerprint positioning and improve the positioning accuracy; it can be widely used to provide indoor positioning services for people.