General Five-Component Scattering Power Decomposition with Unitary Transformation (G5U) of Coherency Matrix

The polarimetric synthetic aperture radar (PolSAR) provides us with a two-by-two scattering matrix data set. The ensemble averaged coherency matrix in an imaging window derived using a scattering matrix has all non-zero elements in its three-by-three matrix. It is a full 3 x 3 matrix that bears nine real-valued and independent polarimetric parameters inside. In the proposed decomposition method, G5U, we preprocess observed coherency matrix [T] by using two consecutive unitary transformations to become an ideal form for five-component decomposition. The transformation reduces nine parameters to seven, which is the best fit for five-component scattering model expansion. We can retrieve five powers corresponding to surface scattering, double bounce scattering, volume scattering, oriented dipole scattering, and compound dipole scattering, directly. These powers can be calculated easily and used to display superb polarimetric RBG images as never before, and are further applicable for polarimetric calibration, classification, validation, etc.

Automated summary

The polarimetric synthetic aperture radar (PolSAR) provides a complete 3x3 matrix with nine real-valued and independent polarimetric parameters. In the proposed decomposition method, the observed coherency matrix [T] is preprocessed using two consecutive unitary transformations to fit a five-component scattering model. This allows for the retrieval of five powers corresponding to different types of scattering, which can be used for various applications such as polarimetric calibration and classification.