Library navigation system implementation method based on deep learning and augmented reality

The invention relates to a library navigation system implementation method based on deep learning and augmented reality. The system implementation method comprises the following steps of: 1, generating an indoor navigation map according to a planar structure of an indoor place; 2, selecting a plurality of places indoors as starting points of navigation, placing different pictures as marked graphs,and inputting the pictures and corresponding coordinate positions into a database; 3, constructing a system database, creating a database required by the system, and updating the database; 4, scanning the identification identifier image, and determining an initial position; and 5, superposing the virtual navigation route with a picture captured by the camera in real time according to the route result obtained by calculation in the above steps, and displaying the virtual navigation route and the picture on a screen of the mobile phone terminal. The indoor positioning and navigation method is high in precision and can