Autofocus compressed sensing imaging based on nonlinear conjugate gradient

Abstract

In this paper, the autofocus compressed sensing (ACS) imaging method is proposed to obtain the imagery with some unknown parameters in the forward model. The proposed method updates the unknown parameters of the measurement matrix and constructs the imagery alternately within an iterative framework. The unknown parameters, denoted as hyperparameters, are estimated using the nonlinear conjugate gradient method. The proposed ACS imaging method is validated using through-the-wall imaging radar data.