Automatic Extraction Method of Urban Road Curb Boundary from Vehicle-Borne Laser Point Clouds

With the acceleration of urbanization, urban road networks are being extended and updated at an alarming rate. How to extract the boundary information of the urban road network efficiently and scientifically to support fine urban management has become a difficult problem. Vehicle-Borne Mobile Mapping System can quickly and efficiently obtain 3D road data, and further realize the extraction of urban road boundary. Based on the tracking information of the Vehicle-Borne Mobile Mapping System, this paper extracts contiguous laser scanning lines from vehicle-borne mobile laser scanning data. By further analyzing the spatial geometric distribution features of the point clouds on different terrain of scan lines, this paper proposes an adaptive window clustering classification method based on the scan line index to solve the problem of the automatic extraction of urban road curb boundary. Experiments are carried out on two point clouds data obtained by the vehicle-borne mobile measurement system. And the accuracy of the extraction of curbs boundary is 99%. The experiment results show that this method can effectively reduce the interference of road noise points as well as the influence of facade features on the boundary extraction of curbs and adapt to different shapes and distribution conditions of urban roads curbs.