Gridless sparse ISAR imaging using modified atomic norm minimization

This paper is concerned with high resolution ISAR imaging via continuous compressed sensing (CCS). Traditional compressed sensing (CS) methods suffer from basic mismatches when off-grid scatters exist. Herein, based on the well-known CCS method i.e. atomic norm minimization (ANM) which deals with the off-gird spectral estimation problem but has a resolution limit, we propose a modified atomic norm algorithm named MANM to achieve a higher resolution and extend its application to ISAR imaging. The new technique adopts a non-convex arctangent sparse metric to promote sparsity more strongly than the existing atomic ℓ₁ norm. The satisfying imaging performance of the proposed method is demonstrated by simulated and experimental results.