A Boosting SAR Image Despeckling Method Based on Non-Local Weighted Group Low-Rank Representation

Area of Science:

• Remote Sensing
• Image Processing
• Signal Processing

Background:

• Synthetic Aperture Radar (SAR) images are susceptible to noise, which degrades image quality and hinders interpretation.
• Existing despeckling methods often struggle to balance noise reduction with the preservation of fine image details and textures.

Purpose of the Study:

• To propose a novel and effective boosting synthetic aperture radar (SAR) image despeckling method.
• To enhance the performance of low-rank representation (LRR) based denoising techniques for SAR images.
• To improve both objective performance metrics and subjective visual quality of despeckled SAR images.

Main Methods:

• A non-local weighted group low-rank representation (WGLRR) model is proposed for SAR image despeckling.
• The method leverages the low-rank property of grouped similar patches from SAR images.
• Pixel corruption probability is integrated as weights to constrain the fidelity of recovered noise-free patches, and a weighted averaging procedure aggregates patch estimations.

Main Results:

• The proposed WGLRR method demonstrates superior performance compared to existing techniques on both simulated and real SAR images.
• Objective evaluation metrics show significant improvements in noise reduction and detail preservation.
• Subjective visual assessment confirms enhanced perceived image quality after applying the WGLRR despeckling method.

Conclusions:

• The WGLRR method offers an effective approach for SAR image despeckling, outperforming conventional techniques.
• The integration of spatial structure, low-rank properties, and pixel fidelity constraints leads to robust noise removal.
• This method provides a valuable tool for improving the quality and interpretability of SAR imagery in various applications.