Vehicle positioning method in viaduct based on map features fused by multiple sensors

The invention discloses a vehicle positioning method in a viaduct based on map features of multi-sensor fusion, and the method specifically comprises the following steps: S1, carrying out the pre-positioning through employing a vehicle-mounted GPS, carrying out the extraction of laser radar feature points, and obtaining laser radar data; S2, adopting a visual camera to obtain a visual positioning image, after a pixel space coordinate system is established, carrying out feature point extraction on the visual positioning image, then converting extracted feature point coordinates into a world coordinate system, and obtaining visual image data; s3, matching and fusing the laser radar data in the step S1 and the visual image data in the step S2 to obtain matched feature points and a corrected perception data set; s4, matching the matched feature points obtained in the step S3 with a map to obtain travel feature points on the map and a travel feature point data set; and S5, calculating the location of the vehicle a