The Fusion of Morphological and Contextual Information for Building Detection from Very High-Resolution SAR Images

Nowadays, very high-resolution synthetic aperture radar (VHR SAR) images are available for interpretation of the built-up area. Buildings are one of the most important parts of the urban area, and in this paper a new method for building detection from a single SAR image is proposed. First, the contextual information of the buildings, such as double bounce, layover and shadow areas are extracted. Using these features, a set of primary detection is made. Second, morphological profiles (MP), with different structural elements (SE), are utilized to build a differential morphological profile (DMP) that provides the building structural information. This structural information is used to make a secondary detection set of building candidates. The final detection result is made by fusion of these two sets. Performance evaluation of the proposed method is reported by the implementation of the method on two different real TerraSAR-X images. The results show that the proposed method has a high detection rate (DR), while the false alarm rate (FAR) is low.