SparsityandCompressed创新.PDFVIP

I N V I T E D P A P E R Sparsity and Compressed Sensing in Radar Imaging The success and accuracy of remote sensing with Radar can be predicted from reasonably limited samples of Radar signals. By Lee C. Potter, Senior Member IEEE , Emre Ertin, Member IEEE , Jason T. Parker, Member IEEE , and Mujdat C¸etin, Member IEEE ¨ ABSTRACT | Remote sensing with radar is typically an ill-posed I. INTRODUCTION linear inverse problem: a scene is to be inferred from limited Radar imaging is an inverse scattering problem whereby a measurements of scattered electric fields. Parsimonious models spatial map of reflectivity is reconstructed from measure- provide a compressed representation of the unknown scene and ments of scattered electric fields. Imaging techniques to offer a means for regularizing the inversion task. The emerging exploit parsimony in sparse or compressible scenes have field of compressed sensing combines nonlinear reconstruction been proposed throughout the 60-year development of algorithms and pseudorandom linear measurements to provide radar processing for suppression of sidelobes and super- reconstruction guarantees for sparse solutions to linear inverse resolution of scattering locations. Many radar processing problems. This paper surveys the use of sparse reconstruction tasks can be posed as finding sparse solutions to under- algorithms and randomized measurement strategies in radar determined linear equationsVa topic addressed by the processing. Although the two themes have a long history in radar