Classification of interferometric synthetic aperture radar image with deep learning approach

Abstract

Interferometric Synthetic Aperture Radar (In-SAR), an extension and further development of Synthetic Aperture Radar (SAR), is widely used in many fields. The intensity map and coherence map obtained from In-SAR data have a strong correlation in space and time, which can be used for the classification of In-SAR image. However, it is not easy to manually explore their correlation and extract features. In this paper, a classification method for In-SAR image based on deep learning is proposed. The deep belief network (DBN) is used to model In-SAR data, which can fully explore the correlation between intensity and the coherence map in space and time, and extract its effective features. The proposed method is tested by the Radarsat-2 C-band In-SAR data of Phoenix and TerraSAR-X x-band In-SAR data of San Francisco, the experimental results show the validity and accuracy of the method.