An Automatic Extraction Method for Pole-Like Objects fromVehicle-Borne Laser Point Cloud

The pole-like objects are the impontant infrastructure in cities, and their automatic extraction is important for fast updating of smart city data. The vehicle-borne mobile measurement system can quickly acquire high-precision point cloud data of the objects in both roadsides. An automatic extraction method for pole-like objects from vehicle-borne laser point cloud data based on circular arc characteristic is proposed. Firstly, according to the target distribution morphology in single and adjacent scanning lines, the non-pole column part in original point cloud is filtered by clustering and denoising. Then, the circular arc point set is searched by constraint random sample consensus (RANSAC) operator, and the 3D statistical regularity of the circular arc are used to identify the column part of pole-like object accurately. Finally, according to the spatial morphology of the target point cloud, the condition of region growing is dynamically determined, and the complete point cloud of different pole-like objects is searched. Experimental results show that this method can effectively reduce noise interference of non-pole-like objects nearby in pole-like objects extraction and has better adaptive ability.