An augmented reality (AR) client terminal identifies initial location information for a location of a user. A scene picture associated with the location of the user is captured. Image matching is performed between the captured scene picture and scene pictures in a scene picture set. The scene picture set is generated based on the initial location information. Each scene picture in the scene picture set is associated with detailed location information. In response to the captured scene picture matching a particular scene picture in the scene picture set, refined location information for the location of the user is determined based on the detailed location information of the particular scene picture.