Multifrequency Bayesian compressive sensing methods for microwave imaging.

Area of Science:

• Electromagnetics and wave propagation
• Computational electromagnetics
• Inverse problems

Background:

• Sparse scatterer imaging is crucial for applications like radar and remote sensing.
• Multifrequency transverse magnetic (TM) illuminations offer rich spectral information for enhanced resolution.
• Bayesian methods provide a robust framework for solving inverse problems with uncertainties.

Purpose of the Study:

• To develop and evaluate innovative Bayesian imaging strategies for sparse scatterer retrieval using multifrequency TM data.
• To introduce customized single-task and multitask Bayesian compressive sensing algorithms for inverse problem solutions.
• To assess the effectiveness and noise robustness of the proposed techniques against state-of-the-art methods.

Main Methods:

• Formulation of two novel imaging strategies: frequency-hopping multistep and multifrequency one-shot schemes.
• Implementation of customized single-task and multitask Bayesian compressive sensing (CS) algorithms.
• Processing of spectral content from microwave scattering data.

Main Results:

• Demonstrated effectiveness of the proposed Bayesian compressive sensing techniques in sparse scatterer imaging.
• Quantified robustness against noise for both frequency-hopping and one-shot imaging schemes.
• Comparative analysis showing advantages over existing deterministic strategies.

Conclusions:

• The developed Bayesian imaging strategies offer a powerful approach for sparse scatterer retrieval.
• The customized Bayesian compressive sensing implementations provide effective solutions for inverse scattering problems.
• The proposed techniques show significant potential for improving radar imaging and related applications.