1. INTRODUCTION
Compared to single-polarimetric SAR images, fully polarimetric SAR contains a wealth of scattering information and exhibits unique imaging characteristics. Consequently, fully polarimetric SAR finds extensive applications in military reconnaissance, forest land development, and disaster detection. PolSAR image classification, a crucial component of remote sensing image processing, has attracted significant attention from researchers. In recent decades, both traditional methods and deep learning techniques have made substantial advancements in the task of PolSAR image classification.