Near-field sparse SAR imaging based on multi-constraint optimization

Abstract Synthetic aperture radar (SAR) imaging is crucial for radar cross section (RCS) measurement. In near-field RCS measurement with growing demand for rapid, convenient testing, it faces challenges of large data volume and high computational complexity. Compressed sensing is adopted to reduce sampling rates below the Nyquist limit, alleviate data burdens, and improve efficiency and imaging performance. This paper proposes a near-field sparse SAR imaging method based on multi-constraint optimization, integrating high-order total variation (HOTV) regularization, exponential regularization, hybrid sparsity constraints, and a time–frequency domain constraint. HOTV introduces a second-order gradient norm to preserve details, suppress noise, and enhance target features, outperforming first-order models. Exponential regularization adaptively adjusts gradient penalty strength, balancing detail preservation and noise suppression. The time–frequency domain constraint further improves imaging quality, while an approximate observation operator reduces computational cost significantly. The optimization problem is solved via alternating direction method of multiplier. Simulation and real-data experiments validate the algorithm’s superior image quality and effective noise suppression, making it suitable for near-field RCS measurement.