ISAR Imaging Based on MEMP Method and Low Rank Matrix Denoising Technique

Abstract

Matrix enhancement and matrix pencil (MEMP) method is an efficient approach for 2-D frequencies estimation in ISAR imaging. However, the MEMP method belongs to the category of spectral estimation algorithm and is sensitive to noise. In the MEMP method, an enhanced matrix formed using the echo matrix is leveraged to extract the 2-D frequencies. Based on the low rank property of this enhanced matrix, we propose a low rank matrix denoising technique for MEMP method. Simulation and quasi real data results demonstrate the effectiveness and robustness of the proposed denoising approach. With the proposed denoising technique, the tolerance to noise of MEMP method has been improved by 10dB.