High-rise building detection in dense urban area based on high resolution SAR images

Abstract

Building extraction is an important task for SAR applications in many fields, such as urban planning, population distribution, and environment evaluation. In this paper, we focus on high-rise buildings in complex urban scenarios. The specific characteristics of high-rise building in high resolution (HR) SAR images are analyzed firstly. In particular, we explore the contextual information of bright line features as corner lines and double bounce; extract the textural feature of the periodic patterns appearing in the layover area. Given those signature details, a new algorithm exploit the building information is proposed for high-rise building detection from single SAR images. The algorithm is applied on spotlight TerraSAR-X images of the center districts in Toronto, Canada, which confirm the applicability of the method for variable high-rise buildings.