A METHOD OF URBAN ROAD NETWORK EXTRACTION BASED ON FLOATING CAR TRAJECTORY DATA

Abstract. With the gradual opening of floating car trajectory data, it is possible to extract road network information from it. Currently, most road network extraction algorithms use unified thresholds to ignore the density difference of trajectory data, and only consider the trajectory shape without considering the direction of the trajectory, which seriously affects the geometric precision and topological accuracy of their results. Therefore, an adaptive radius centroid drift clustering method is proposed in this paper, which can automatically adjust clustering parameters according to the track density and the road width, using trajectory direction to complete the topological connection of roads. The algorithm is verified by the floating car trajectory data of a day in Futian District, Shenzhen. The experimental results are qualitatively and quantitatively analyzed with ones of the other two methods. It indicates that the road network data extracted by this algorithm has a significant improvement in geometric precision and topological accuracy, and which is suitable for big data processing.