A Nonlocal TV-Based Variational Method for PolSAR Data Speckle Reduction.

Area of Science:

• Remote Sensing
• Image Processing
• Signal Processing

Background:

• Speckle noise significantly degrades the quality of polarimetric synthetic aperture radar (PolSAR) data.
• Existing speckle reduction methods often struggle to preserve fine details and polarimetric scattering characteristics.

Purpose of the Study:

• To propose a novel variational model for effective speckle reduction in PolSAR data.
• To enhance the preservation of image details, textures, edges, and polarimetric scattering properties.

Main Methods:

• A nonlocal total variation (NLTV)-based variational model named WisNLTV is developed.
• The model incorporates a Wishart fidelity term and NLTV regularization for complex-valued fourth-order tensor data.
• An efficient bi-convex model and an alternating minimization algorithm are employed for solving the non-convex problem.

Main Results:

• The proposed iteration algorithm demonstrates theoretical and numerical convergence.
• Experimental results show superior speckle reduction compared to other methods.
• The WisNLTV method effectively preserves image details, repetitive structures, and polarimetric scattering characteristics.

Conclusions:

• The WisNLTV model offers a robust and effective solution for PolSAR speckle reduction.
• The method achieves a superior balance between noise suppression and feature preservation.
• WisNLTV advances the state-of-the-art in PolSAR image processing.