High resolution sar image analysis by combining the pca and the azimuth sub-band decompositions

Kurzfassung

With the increase of the Synthetic Aperture Radar (SAR) sen- sor resolution, a more detailed analysis and a ¯ner description of SAR images are needed. However, in the case of urban ar- eas, the high diversity of man-made structures combined with the complexity of the scattering processes make the analysis and information extraction from high resolution SAR images non- trivial. Moreover, the need for automatic processing and inter- pretation of large-size SAR data volumes has become more and more pressing. Therefore, new innovative tools are required for a better assessment and a ¯ner description. This paper proposes an automatic SAR image interpretation approach based on the marriage of two techniques: the Principal Components Analysis (PCA) and the azimuth sub-band decom- position. Indeed, both methods were proved to be of great in- terest in the analysis and modeling of the SAR image contents. Thus, a combination should be more informative and better de- scriptive. In this article, both a description of two combination ways that we propose to use, and a presentation of our prelimi- nary performance results will be provided.