Automated RCS Measurements Processing via Multibasis Dictionary and Compressive Sensing

Accurate extraction of radar cross section (RCS) from real-world measurements is challenged by noise, clutter, and near-field effects. This article presents the dictionary pursuit (DP) method, a framework based on compressive sensing that addresses these issues. The DP method employs distinct image and range-Doppler bases to simultaneously represent the target and environmental contamination within a single, weighted L1-norm optimization. Building upon previous concepts explored by other authors, this work provides a detailed mathematical formulation, algorithmic improvements, and a comprehensive workflow for this multibasis approach. Through validation with both numerical simulations and experimental measurements, we demonstrate the DP method’s enhanced performance compared to classical Fourier-based techniques. It automatically separates the target signature from clutter and noise, achieves a more accurate near-field to far-field transformation (NFFFT) by avoiding approximate inverse transforms, and inherently yields a higher quality super-resolved inverse synthetic aperture radar (ISAR) image with reduced sidelobes.