Automatic Extrapolation of Missing Road Network Data in OpenStreetMap

Stefan Funke, Robin Schirrmeister, Sabine Storandt

Abstract: Road network data from OpenStreetMap (OSM) is the basis of various real-world applications such as fleet management or traffic flow estimation, and has become a standard dataset for research on route planning and related subjects. The quality of such applications and conclusiveness of research crucially relies on correctness and completeness of the underlying road network data. We introduce methods for automatic detection of gaps in the road network and extrapolation of missing street names by learning topological and semantic characteristics of road networks. Our experiments show that with the help of the learned data, the quality of the OSM road network data can indeed be improved.