A Three-Dimension Polarimetric Correlation Pattern Interpretation Tool and Its Application

A new three-dimension polarimetric correlation pattern (3-D PCP) is developed in this paper by extending the polarimetric correlation pattern along the polarimetric ellipti-city angle. Three representative 3-D correlation patterns can be thus derived. Then, the periodicity and extreme points distribution are analyzed for four canonical scatterers in detail. Signifi-cant differences can be observed in their manifold pattern structures and several physical meaningful features including the Gaussian curvature in differential geometry are extracted to an-alyze the polarimetric response with the change of polarimetric rotation angle and polarimetric ellipse angle. On this basis, the ship and sea pixels of the Radarsat-2 PolSAR data are analyzed by using the three-dimensional polarimetric correlation pattern of HH and HV channels. The ship region can be well extracted utilizing the Gaussian curvature feature, which shows the poten-tial in ship detection. The results prove that the three-dimensional polarimetric correlation pattern and the extracted features have the ability to distinguish different scattering mecha-nisms.