2-D Autofocus for High-Squint SAR Based on Affine Coordinate Back-Projection Algorithm
Yishan Lou,Hao Lin,Quan Chen,Mengdao Xing,Shengwei Zhou,Guang-Cai Sun
DOI: https://doi.org/10.1109/tgrs.2024.3493095
IF: 8.2
2024-01-01
IEEE Transactions on Geoscience and Remote Sensing
Abstract:When synthetic aperture radar (SAR) works in high squint (HS) mode, the interpolation and scaling operation of traditional frequency domain imaging algorithms will change the original structure of motion error, and lead to imaging difficulties. As a classical time-domain imaging algorithm, the back-projection (BP) algorithm is linear processing with a high tolerance for motion error. Therefore, the BP algorithm is very suitable for HS SAR imaging. To further improve the estimation accuracy of motion error, an innovative affine coordinate (AC) system is introduced into the BP algorithm. Based on this AC system, a novel two-dimensional (2-D) autofocus algorithm is proposed, which can more accurately estimate and correct the 2-D phase error of the HS SAR BP image. The proposed algorithm has the following advantages: i) The AC imaging grid is established according to the proposed resolution calculation method based on the BP image spectrum. Under this imaging grid, the nonsystem range cell migration (NsRCM) and the range defocus term of the BP image are significantly reduced, making the phase error estimation more accurate. ii) A spectrum alignment processing for BP image in the AC system is proposed to remove the spectrum aliasing so that the azimuth phase error (APE) can be accurately estimated. iii) The spectrum of the BP image is orthogonal in the AC system, which makes the 2-D phase error compensation based on the established prior phase error structure more accurate. Simulation and real data experiments validate the performance of the proposed algorithm.