Statistical Modeling of the Reflection Symmetry Metric for Sea Clutter in Dual-Polarimetric SAR Data

摘要

Reflection symmetry is an intrinsic property used to distinguish man-made metallic targets and sea clutter in dual-polarimetric synthetic aperture radar (SAR) data. We derive a statistical model for the reflection symmetry metric (RSM) of sea clutter using the covariance matrix between co-polarized and cross-polarized channels. Consequently, the model is theoretically analyzed for its appropriateness in describing the statistical behavior of the RSM of sea clutter. We also assess the suitability of the Student’s T location-scale (TLS) distribution for modeling the RSMs of sea clutter in dual-polarimetric SAR data. Experiments based on single look complex (SLC) HH–HV and VV–VH dual-polarimetric SAR data acquired by L-band ALOS–PALSAR and C-band RADARSAT-2 verify the superiority of the model over the TLS distribution.