Satellite SAR and Human Settlement Detection

Abstract

Due to the peculiar features of human settlement areas, with enormous variance of objects and the resulting variability and possibly also ambiguity of data, such areas represent the most demanding cases for information extraction from SAR data. This research is particularly aimed at human settlement mapping at a regional scale. Indeed, informal settlement monitoring is an important topic for the many national and international initiatives including GMES and the humanitarian and development aid policies of major countries. Also, monitoring settlements is related to phenomena like illegal immigration that are very high on the list of policy makers; finally, SAR is still underestimated as a reliable source of data. From a technical point of view, this work is based on the integration of spatial information extracted from SAR data to reach a suitable mapping accuracy for urban areas. In particular, the use of co-occurrence textural features has been considered, and the role of the spatial scales in the images for determining different environment investigated. In particular, the procedure will be discussed for the Chinese region of Wuhan using ENVISAT SAR data.