Abstract: Accurate identification of main roads and auxiliary roads is essential for the automatic synthesis of complex intersections. Traditional methods mostly rely on the straightness and curvature of road sections to describe the main roads and auxiliary roads at intersections. However, complex intersections have intricate structures and various morphological changes. Reliance on these two features for the identification of main and auxiliary roads is limited, and some forms Similar main branches cannot be identified. Therefore, this paper proposes a method for identifying main and auxiliary roads at complex intersections with multiple feature constraints. First extract feature points of complex intersections, interrupt arc segments based on feature points, identify parallel arc segments based on the straightness of the road segments, and then identify parallel clusters based on road ductility, angle, distance and other features to obtain complex intersections. Arterial roads, followed by identifying ramps on compactness and distance relationships to obtain auxiliary roads at complex intersections. Taking Nanjing OSM urban road network as an example, experiments show that the method in this paper can accurately identify the main road and auxiliary road sections of complex intersections, and the recognition accuracy is 93.60% and 89.43%, respectively.

Key words: complex intersections, identification of main and auxiliary roads, feature points, parallel cluster, ramp