Visual positioning method and map construction method and device

The invention discloses a visual positioning method. The method comprises the following steps: loading a map constructed based on a ground image; collecting a current ground image and current inertial sensor data at any moment during traveling, extracting image feature points based on a current ground image frame, and calculating matching information of the image feature points; searching map nodes based on a map according to the current inertial sensor data; and performing map point position information and matching information matching on the image feature points and each map point in the map nodes, and if the matching succeeds, calculating the global pose of the current image frame according to the matched feature points to obtain a current positioning result. According to the invention, mapping and visual positioning are carried out by using the ground image so that the dependence on manual marking is reduced. 公开了一种视觉定位方法，包括，加载基于地面图像构建的地图；在行进中的任一时刻采集当前地面图像和当前惯性传感器数据，基于当前地面图像帧，提取图像特征点，计算图像特征点的匹配信息；根据当前惯性传感