Synthetic aperture radar imaging and motion estimation via robust principal component analysis (Q2873247)

The paper deals with the problem of synthetic aperture radar (SAR) imaging of scenes with multiple targets: many stationary targets and a few moving targets. The SAR setup is the usual one with a single moving antenna emitting and receiving signals. The authors focus on the detection of the moving targets and the separation of the echoes from the stationary targets and those from the moving targets. After such separation the stationary scene can be imaged by itself, and the motion estimation can be carried out on the echoes from the moving targets alone. The authors propose and analyze a detection and data separation approach based on the robust principle component analysis (robust PCA) method. Quoting from the Introduction: ``The main contribution of this paper is to show with analysis and numerical simulations that by appropriately pre-processing and windowing the SAR data we can decompose it into a low rank part, corresponding to the stationary scene, and a sparse part, corresponding to the moving targets. Our theoretical and numerical study describes the rank of the pre-processed SAR data as a function of the velocity, location, and density of the scatterers. It specifies in particular how slowly a target can move and still be distinguishable from the stationary scene. It also addresses the question of proper windowing of the data for the separation with robust PCA to work.'' The paper includes numerical simulations illustrating that robust PCA, complemented with proper data windowing, can be used for motion detection and data separation.