A road-network matching approach guided by 'structure'

With the increasing availability of diverse geospatial databases over more than two decades, the road-network matching has become an issue of growing interest. To a certain extent, a road network can be regarded as a unit characterized by various reoccurring structures, such as dual carriageways (parallel lines), roundabouts, narrow passages, navigation stubbles and slip roads around cloverleaf junctions. As the various road structures take on different geometric or topologic characteristics, it is hardly possible to efficiently match all of them using the same criteria or methods. To circumvent this problem, a new matching approach guided by 'structure' is proposed in this article, which is composed of three consecutive processes: category recognition, process modelling and process execution. The structure-guided matching approach is able to classify the road network into several structure categories and then assign to each matching category a specific matching strategy with proper parameters or criteria, which allows a context-related analysis and thus considerably improves the matching performance in terms of matching rate and accuracy.