Subway rail obstacle detection method fusing vision and laser radar data features

The invention discloses a subway rail obstacle detection method fusing vision and laser radar data features. The method comprises the steps of s1, assembling a laser radar and a camera for a subway train; s2, driving the subway train assembled with the laser radar and the camera to travel for at least one time along a driving line, acquiring camera image data of the line and laser radar point cloud data; s3, constructing a map by utilizing the laser radar point cloud data; calibrating the positions of crossed tracks and the position of curves one by one by utilizing the image data of the camera; s4,manually marking the condition that cannot be determined through the camera image data and the laser radar point cloud data to form a subway map data of the driving line in s2. According to themethod, the visual and laser radar data features are fused, and the barrier detection of a subway track is realized; the laser radar data and vision data are mined with the laser radar data as the main data, and the visual fe