Clutter statistics in high-resolution synthetic aperture radar imagery

This paper studies the applications of squared radius and trace statistic under homogeneous and texture models for modeling terrain radar clutter in polarimetric synthetic aperture radar (SAR) data. The squared radius and trace statistic involve the use of scattering vector and polarimetric covariance matrix, respectively, in their computation. The terrain radar clutter modeling was performed on TerraSAR-X single-look high-resolution spotlight data and multi-look NASA/JPL AIRSAR POLSAR data. Both the homogeneous and texture models were investigated and evaluated quantitatively through chi-squared goodness-of-fit test. The obtained results showed that the texture model provided a better fitting compared with the homogeneous model.