An approach of building areas segmentation of SAR images based on the level set method

Abstract

Because of the imaging characteristic of Synthetic Aperture Radar (SAR) building areas, an approach based on the level set approach is used to segment building areas. We use the edge feature based on exponential wavelet and texture feature based on variogram to build the energy criterion model, and then the evolutive curve is implemented by minimizing the energy criterion model via variational level set approach to detect the building areas. Compared to the traditional energy criterion model based on the gray information, experimental results show that the proposed model is effective and building areas of large SAR image is detected accurately. © Springer-Verlag Berlin Heidelberg 2011.