SAR imaging via modern 2-D spectral estimation methods

Area of Science:

• Remote Sensing
• Signal Processing
• Radar Imaging

Background:

• Conventional Fourier transform methods in synthetic aperture radar (SAR) imaging face limitations in resolution, sidelobe artifacts, and speckle.
• Advanced spectral estimation techniques offer potential improvements for SAR data processing.

Purpose of the Study:

• To comprehensively compare 2D spectral estimation algorithms for SAR imaging.
• To develop and evaluate multichannel variants of adaptive algorithms for interferometric SAR (InSAR) applications.

Main Methods:

• A comparative analysis of various 2D spectral estimation algorithms, including novel derivations and insights.
• Development of multichannel variants for minimum variance method (MVM), reduced-rank MVM (RRMVM), adaptive sidelobe reduction (ASR), and space variant apodization (SVA).
• Application of these algorithms to simulated and collected SAR data, including polarimetric displaced-aperture interferometric data.

Main Results:

• Demonstrated performance of different 2D spectral estimation methods on SAR imagery.
• New multichannel variants of MVM, RRMVM, ASR, and SVA were developed for interferometric height and reflectivity estimation.
• Adaptive spectral estimation improved height estimates through adaptive nulling and averaging in interferometric SAR.

Conclusions:

• MVM, ASR, and SVA provide significant advantages over Fourier methods for SAR imaging, improving both scattering intensity and interferometric height estimation.
• The developed interferometric variants offer enhanced accuracy for height estimation in InSAR.
• This work provides a comprehensive comparison and novel contributions to adaptive SAR imaging techniques.