[Paper Review] Fast 3D Synthetic Aperture Radar Imaging from Polarization-Diverse Measurements

Pierre Minvielle, Pierre Massaloux+1 more|arXiv (Cornell University)|Jun 24, 2015

Synthetic Aperture Radar (SAR) Applications and Techniques20 references3 citations

TL;DR

This paper presents a fast, regularized 3D radar imaging method that simultaneously reconstructs three high-resolution 3D scatterer maps (HH, VV, HV) from polarization-diverse synthetic aperture radar (SAR) measurements. By extending the multiple scattering model into a high-dimensional linear inverse problem and solving it via a dedicated fast algorithm, the method enables accurate identification and characterization of complex radar backscattering components, validated on real and simulated data with strong agreement between synthetic and measured results.

ABSTRACT

An innovative 3-D radar imaging technique is developed for fast and efficient identification and characterization of radar backscattering components of complex objects, when the collected scattered field is made of polarization-diverse measurements. In this context, all the polarimetric information seems irretrievably mixed. A direct model, derived from a simple but original extension of the widespread "multiple scattering model" leads to a high dimensional linear inverse problem. It is solved by a fast dedicated imaging algorithm that performs to determine at a time three huge 3-D scatterer maps which correspond to HH, VV and HV polarizations at emission and reception. It is applied successfully to various mock-ups and data sets collected from an accurate and dedicated 3D spherical experimental layout that provides concentric polarization-diverse RCS measurements.

Motivation & Objective

• To address the challenge of efficiently processing polarization-diverse SAR data for 3D radar imaging in RCS and target signature analysis.
• To overcome the limitations of conventional ad hoc localization techniques that fail to resolve complex scattering components.
• To develop a fast, simultaneous reconstruction method for multiple polarization states (HH, VV, HV) in a single inversion step.
• To enable accurate 3D localization of diffraction, reflection, and depolarizing scatterers—especially corners and edges—critical for target characterization.
• To validate the method on a dedicated 3D spherical measurement setup with high-fidelity data from mock-ups and real targets.

Proposed method

• The method is based on a novel extension of the standard multiple scattering point model, incorporating polarization diversity into a high-dimensional linear inverse problem.
• It formulates the 3D radar imaging problem as a regularized least-squares inversion to recover three 3D scatterer maps (HH, VV, HV) simultaneously.
• A fast, dedicated algorithm solves the large-scale linear system using the Modified Nonlinear Least Squares (MNLS) approach with Tikhonov-type regularization.
• The spherical measurement geometry enables concentric, full-azimuthal, polarization-diverse RCS measurements, improving angular sampling and scattering component resolution.
• The algorithm leverages the polar nature of frequency-domain data and applies efficient Fourier-based processing to reduce computational cost.
• The method is validated using both synthetic data from a 3D electromagnetic solver and real data from a custom 3D spherical anechoic chamber setup.

Experimental results

Research questions

• RQ1 Can polarization-diverse SAR measurements be efficiently processed to reconstruct three independent 3D scatterer maps (HH, VV, HV) in a single, fast inversion?
• RQ2 How does the proposed extension of the multiple scattering model improve the resolution and characterization of complex scattering components such as edges and corners?
• RQ3 To what extent does the method preserve fidelity when reconstructing 3D scatterer maps from real experimental data compared to simulated EM solver results?
• RQ4 What role does polarization diversity play in resolving multiple scattering artifacts and depolarizing scatterers like target corners?
• RQ5 How do near-field effects and angular undersampling impact the reconstruction quality, and can they be mitigated through improved sampling or correction models?

Key findings

• The 3D radar images reconstructed from real data (HH, VV, HV) show strong agreement with those from a parallelized 3D electromagnetic solver, confirming the method's accuracy.
• The method successfully localizes not only primary reflections and diffractions but also depolarizing scatterers such as corners, which are critical for target signature analysis.
• For the arrow mock-up, the HH and HV maps show enhanced sensitivity to multiple reflections between the arrow's base and rear, indicating high resolution in detecting wave interactions.
• The glider mock-up results reveal angular undersampling artifacts in the wing’s trailing edge backscatter, indicating limitations due to transverse spread and insufficient angular sampling.
• The method demonstrates robustness in resolving complex scattering patterns under realistic measurement conditions, with results consistent across both synthetic and real data sets.
• The simultaneous reconstruction of three polarization states enables comprehensive RCS analysis by distinguishing between specular, diffractive, and depolarizing components.