High Resolution Imaging of ISAR Based on Low Rank Matrix Denoising Technique
Download as PDF
DOI: 10.25236/iwmecs.2019.003
Author(s)
Yuanhang Tang, Yang Xu, Yongsheng Zhang
Corresponding Author
Yuanhang Tang
Abstract
The problem of inverse synthetic aperture radar (ISAR) imaging can be expressed as a two-dimensional frequency estimation problem. Matrix enhancement and matrix pencil (MEMP) method is an effective frequency estimation method. This method belongs to the category of spectral estimation methods, and the performance of ISAR imaging using MEMP is sensitive to noise, which limits the application of this method in ISAR imaging. In the MEMP method, an enhanced matrix with a large dimension is first constructed according to the echo matrix. The enhanced matrix has low rank property when the scattering points in the imaging scene satisfy the sparse characteristic. Based on the low rank property of the enhanced matrix, we propose a denoising method for the MEMP. The simulation results prove that the proposed denoising method combining with the MEMP can achieve robust and effective high resolution imaging for ISAR.
Keywords
ISAR Imaging, Frequency Estimation, Matrix Enhancement and Matrix Pencil, Low Rank Matrix Denoising