Markovian graph labeling for 3D reconstruction in dense urban area using SAR and optical images

In this paper, a method to extract a simplified 3D map of dense urban areas is proposed. It relies on both optical and interferometric SAR images. Firstly, an over-detection of the buildings is done over the optical image using morphological operators. They are modeled as rectangle with a height that is retrieved using the interferometric phase. Then, the problem is modeled as a Markov Random Field and optimized using many different features retrieved in both optical and SAR data.