SAR imaging using a modern 2D spectral estimation method

Modern 2D spectral estimation techniques can be used for SAR imaging. The aim of those techniques is to improve the SAR image quality at the expense of an additional processing load. This paper focuses on applying the 2D minimum variance power spectrum estimation method (MVM) to SAR raw data. MVM process improves the image ISLR and reduces speckle noise. MVM was found to be particularly useful in SAR imaging geometries involving steep depression angles. Such geometries are characterized by high clutter interference, leading to poor contrasts between targets and the surrounding clutter. MVM process enhances the target to background contrast in such images. Parallel MVM process was demonstrated at Elta on a processing array. MVM images of 8000 by 8000 pixels are produced there in few minutes. Examples of MVM images are given and compared to conventional FFT based SAR images. Quantitative measures for ISLR improvement and speckle smoothening are given.