Similarity among Geographic Networks

International audienceThe available amount of geographic datasets has considerably grown. These data (maps) are available in different formats, in different scales, and they are usually generated by different procedures. Distinct maps that represent the same or overlapping areas (multiresolution maps) can differ both in accuracy and resolution. Therefore, an important issue is to determine whether two maps are consistent, i.e., do they represent the same area without contradictions, and if not, are they at least similar? In this paper we develop a method to assess the similarity over complex structured spatial objects that form networks. Existing approaches deal only with the change in accuracy and do not take into account the change in resolution; as a result the two maps in question must have the same number of components. In this paper we extend them to treat the change in resolution by tying each component in the map with weight according to its importance. Other improvements to existing approaches, which are based on topological properties, are proposed by considering the directional and metrical properties. This method is the first step to assess the similarity between maps with complex configurations including all the geographic features