Speckle reduction in synthetic-aperture-radar imagery

Area of Science:

• Remote Sensing
• Image Processing
• Signal Processing

Background:

• Speckle noise in synthetic-aperture-radar (SAR) images degrades image quality and information content.
• Effective speckle suppression is crucial for accurate SAR image analysis.

Purpose of the Study:

• To investigate adapted local processing techniques for speckle noise suppression in SAR images.
• To evaluate two distinct processing algorithms for their effectiveness in reducing speckle.

Main Methods:

• Development of speckle noise suppression algorithms based on a statistical model accounting for multilook correlation.
• Application of two algorithms: one using image intensity, the other using a homomorphic transformation of intensity.
• Processing of Seasat-A SAR images to test the algorithms.

Main Results:

• Adapted local processing effectively suppresses speckle noise in SAR images.
• Both intensity-based and homomorphic transformation-based algorithms demonstrate speckle reduction capabilities.
• Processed images show improved definition of statistical parameters within local windows.

Conclusions:

• Local processing techniques offer a viable solution for mitigating speckle noise in SAR imagery.
• The developed algorithms enhance the utility of SAR data by improving image interpretability.
• Further experimental validation on Seasat-A data confirms the effectiveness of the proposed methods.