A Decoupled Chirp Scaling Algorithm for High-Squint SAR Data Imaging

The capability to work on the high-squint mode brings great advantages to the synthetic aperture radar (SAR) system. However, the efficient imaging of the high-squint SAR data has not been well settled. Due to the severe range-azimuth coupling (RAC), the conventional imaging algorithms fail to work properly in high-squint cases. In this article, a decoupled chirp scaling algorithm (DCSA) is proposed for high-squint SAR data imaging, whose key step is a range-azimuth decoupling (RAD) preprocessing in the 2-D frequency domain. After RAD, the majority of the RAC is eliminated, where the high-squint SAR data are simplified into the quasi-broadside mode. The data can be then processed via the chirp-scaling-typed scheme, which is summarized as follows. The data after RAD are first transformed into the range-Doppler (RD) domain, in which a nonlinear chirp scaling (NLCS) is performed in the range direction to eliminate the chirp rate variations caused by the residual RAC. The chirp scaling processing is simultaneously performed to equalize the range cell migrations (RCMs) for the full scene data. The range compression, bulk RCM correction, and compensation for the cubic and quartic phases introduced by the NLCS are then accomplished in the 2-D frequency domain. Finally, the azimuth compensation is accomplished in the RD domain to obtain the final focused SAR image. Different from the linear range walk correction (LRWC) accomplished in the azimuth time domain, the RAD method proposed here does not destroy the azimuth-invariant property of the SAR echoes. Therefore, the DCSA can adapt to the wide swath SAR data imaging. Simulation shows that, for the airborne SAR with 1-m resolution in both range and azimuth directions, and a squint angle of 45 degrees, the DCSA can achieve the bulk focusing of the full-scene data with a range swath of 10 km, whereas the azimuth swath is not limited. Compared with the classical CSA, the DCSA requires only one more range fast Fourier transform (FFT)/inverse FFT (IFFT) and one more complex matrix multiplication, which is, therefore, timely and very efficient.