Fast ISAR imaging based on sparse fourier transform algorithm

Range-Doppler (RD) algorithm is conventionally a favorable choice in inverse synthetic aperture radars to generate high-resolution images of the uniformly rotating targets. However, when waveforms with a large time-bandwidth product, such as the linear frequency modulated signals, are adopted to achieve a long-range detection with high resolution, the resulted computational complexity in the imaging process becomes unbearable. Taking into account the fact that the scattering points generally exhibit sparsity in both range and azimuth dimension, we present a fast RD imaging algorithm based on sparse Fourier Transform in this paper. The echo signals from the scattering points are processed in a two-dimensional manner, and a novel coherent integration method is developed to enhance the algorithm robustness in low signal-to-noise ratio scenarios. The overall computational complexity is significantly reduced without compromising the imaging resolution. The effectiveness of the proposed algorithm is validated by numerical simulation results.