A Review on PolSAR Decompositions for Feature Extraction

Area of Science:

• Remote Sensing
• Geospatial Analysis
• Signal Processing

Background:

• Feature extraction is crucial for analyzing complex remote sensing data, particularly fully polarimetric imagery.
• Polarimetric decomposition techniques are essential for extracting detailed information from these datasets.
• Understanding different decomposition methods is key to advancing remote sensing applications.

Purpose of the Study:

• To systematically review and present established polarimetric decomposition algorithms for remote sensing.
• To elucidate the mathematical foundations and principles of coherent and non-coherent decomposition methods.
• To provide insights into potential combinations of these techniques for diverse applications.

Main Methods:

• Categorization of polarimetric decomposition techniques into coherent and non-coherent methods.
• Detailed examination of foundational algorithms: Pauli decomposition.
• In-depth analysis of coherent methods (Cameron) and non-coherent methods (Freeman-Durden, Yamaguchi, Cloude-Pottier).

Main Results:

• Experimental testing of each decomposition method on a benchmark Vancouver area dataset.
• Comparative analysis of the efficacy of various polarimetric decomposition techniques.
• Identification of strengths and underlying principles of each method.

Conclusions:

• Polarimetric decomposition is vital for effective remote sensing data analysis.
• A comprehensive understanding of diverse decomposition techniques enhances feature extraction capabilities.
• Further research into combining methods can unlock new applications in remote sensing.