Abstract: With the development of self-driving high-precision map, it is necessary to extract the road rules accurately and efficiently. This paper takes a city expressway in suzhou city as the research object, obtains the city road point cloud data by using the leica vehicle-borne laser mobile measurement system, observes and understands all kinds of attribute information of the vehicle-borne point cloud data. According to the observation results, the point cloud data height difference, the point cloud gray level difference and the point cloud intensity difference are combined. Finally, the accuracy of the extracted vectorization results is evaluated. The experimental results show that: This method can get good results in extracting regular polygons and can meet the demand of high precision maps.

Key words: high precision map, vehicle point cloud, road rule sign marking, mesh density, 3D vector