Extraction and Update of Street Networks in Urban Areas from High Resolution Satellite Images

Among possibilities of remote sensing for urban uses, extraction of street networks is an important topic. A good knowledge of the street network is necessary for urban planning, map updating or estimation of atmospheric pollution. Moreover, an assisted street extraction from remotely sensed data can highly help cities which do not have a cartography of their road network. However, if road extraction from rural or semi-rural area images has led to encouraging results, the urban context, because of its complexity and the numerous artefacts encountered, is still a new and challenging topic. The context of our current work is urban area and high spatial resolution imagery. In order to help image interpreters, we are concerned about extracting and updating street networks. In this paper, we present a complete methodology adapted to those aims. It includes the definition of a reference object for quantitative evaluation and also the processing algorithm. The extraction process is then applied on a high resolution IKONOS image. Results and future prospects are finally discussed