Segment-based urban block outlining in high-resolution SAR images

Abstract

While analysing remotely sensed images of urban areas, in many cases an at least approximate knowledge of block partition is useful for specialising operations over areas within which a certain degree of homogeneity can be assumed. Unfortunately, though, this information is not always accessible in a reasonable time or with a reasonable effort, or even in some cases a GIS of the city is not available at all. Automatic extraction of block boundaries becomes thus an interesting means of obtaining such information. City blocks are usually separated by major urban roads; this paper presents a preliminary work on the use of a linear feature extractor, originally developed for road network extraction, as a tool to partition a very high resolution SAR scene acquired over an urban area.