GEO SAR Imaging of Maneuvering Ships Based on Time–frequency Features Extraction
Jindong Yu, Ze Yu, Chunsheng Li
Abstract
Geosynchronous synthetic aperture radar (GEO SAR) system has the ability of observing targets for several hours in a large field of view. Complex and irregular rotating motion of the ship on sea in long aperture time makes it difficult to focus. To accomplish the imaging of maneuvering ships, a novel imaging algorithm is proposed in this article. A spatial filter constructed by convolutional neural network (CNN) is utilized to extract the time–frequency features. Combined with local generalized Radon–Fourier transform (GRFT) and phase gradient autofocus (PGA), the extraction accuracy further increases. Validation results based on simulation experiment, airborne experiment, and wave pool experiment show that the proposed algorithm achieves good imaging performance.