期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2010
卷号：XXXVIII Part 4
出版社：Copernicus Publications
摘要：The registration of imagery to maps is becoming increasingly important for a large number of applications. Today, this task still reliesmostly on an operator, manually identifying corresponding points from the image and the GIS dataset for registration. The challenges forautomating the process arise mainly from their dissimilar data structures (raster vs. vector), orientation variations, as well as occlusionintroducedextraction errors due to shadows, buildings or objects on the roads. In this paper, we present a robust automated approach thatmodels road networks extracted from the two datasets as graphs, using statistical and structural descriptions of these road networks. Theproposed approach starts by statistically analyzing local geometrical and topological properties of road networks such as orientation andnumber of connections. Such statistical similarity measures can be used to thin the number of potential matches when comparing a targetroad structure to a spatial database, resulting in computational efficiency. Subsequently, by considering the spatial distribution andstructure similarity in a neighbourhood, we formulate a global compatibility function to measure the overall goodness of correspondence.We achieve an optimal matching by finding an optimal morphism that maximizes this compatibility function. The experimental resultsdemonstrate the robustness of our approach
