Improved image reconstruction from sub-apertures of circular spotlight SAR

Abstract

Reconstructing images from SAR returns is usually a time consuming task. In addition it is often desired to use as many returns as possible to achieve better image quality. However, the high computational resources demand by the conventional methods hinders the reconstruction process. In this article, we propose a simple method to reconstruct SAR image that is built upon the back-projection algorithm using multiple sub-aperture imagery to attain both greater processing efficiency and improved image quality. Instead of aggregating all the available pulses into one single image following the backprojection algorithm, our proposed method creates multiple SAR image reconstructions from a relatively small number of pulses to exploit variations in sub-aperture views of the scene. The heterogeneity among these sub-aperture views exhibits an exceptional difference between various objects and presents a reconstruction with much less noise. Our proposed method is evaluated with circular spotlight GOTCHA data sets and it demonstrates much improved computational performance and image quality compared to the conventional back-projection algorithm.

CNN-Based Vehicle Target Recognition with Residual Compensation for Circular SAR Imaging

“…Kou et al carried out a series of studies on geosynchronous CSAR imaging, and analyzed the effects of orbit error, L-band of troposphere and high sidelobe on imaging [29][30][31]. Yuan et al proposed a method for reconstructing SAR images that is built upon the backprojection algorithm using multiple sub-aperture images to attain both greater processing efficiency and improved image quality [32]. Chen et al proposed a processing strategy for the 3D reconstruction of vehicles, which only needs the single-pass single-polarization CSAR data [33].…”