SAR Image Change Detection via Multiple-Window Processing with Structural Similarity.

Area of Science:

• Remote Sensing
• Geospatial Analysis
• Signal Processing

Background:

• Synthetic Aperture Radar (SAR) imagery is crucial for monitoring Earth surface changes.
• Speckle noise in SAR images significantly degrades the quality of change detection results.
• Existing change detection methods often struggle to balance noise reduction and preservation of subtle changes.

Purpose of the Study:

• To propose a novel SAR change detection approach.
• To enhance the quality of coherence images for improved change detection.
• To effectively reduce speckle noise while preserving critical change information.

Main Methods:

• The proposed method utilizes a two-step process: coherence image generation and gamma correction (GC) filtering.
• Coherence image generation is based on multiple-window processing (MWP) combined with the structural similarity index measure (SSIM).
• Gamma correction (GC) filtering is applied to reduce speckle noise, with an optimized order parameter.

Main Results:

• The MWP operation with SSIM demonstrates high sensitivity to intensity similarity between SAR images, yielding high-quality coherence images.
• Optimized GC filtering effectively reduces speckle noise without significant loss of information from changed regions.
• Experimental results validate the proposed method's effectiveness in SAR change detection.

Conclusions:

• The proposed SAR change detection approach effectively integrates SSIM and MWP for superior coherence image generation.
• The method significantly mitigates speckle noise through optimized GC filtering, enhancing change detection accuracy.
• This technique offers a robust solution for accurate and reliable SAR-based monitoring of surface changes.