ABSTRACT
As is well-known, eigenvalue/eigenvector analysis has mathematical universality. The eigenvalues of the matrix represent scattering powers. The largest eigenvalue has the largest power, and the smallest eigenvalue corresponds to the smallest power. Hence, there is a one-to-one correspondence between the magnitude of eigenvalues and the power itself. If we choose the largest eigenvalue in a PolSAR image, it always shows the largest power throughout the image regardless of the scattering mechanism. Therefore, combinations of eigenvalues, such as H, mean alpha angle, A, and TP are preferred rather than using a single eigenvalue.
