Long-Time Observation-Based Multiattitude ISAR Imaging Method for Moving Ships

Abstract

Long-time synthetic aperture radar observation facilitates the tracking and identification of moving targets. However, continuous changes in the attitude of a moving target during long-duration observations cause the azimuth signal’s time–frequency (TF) trajectory to curve. The curvature of the TF trajectory leads to image defocusing. The varying degrees of trajectory curvature for each scatterer caused by attitude changes prevent the traditional inverse synthetic aperture radar (ISAR) autofocusing method from achieving ideal focusing results. To address this issue, we propose an innovative ISAR imaging method based on TF trajectory extraction and compensation. This method divides the observation into subtime intervals for imaging, allowing the capture of various motion attitudes of the target. By transforming the TF trajectories with varying curvature into horizontal trajectories, the proposed method effectively handles nonstationary intervals, enabling attitude image acquisition during these periods. Experimental results demonstrate that our algorithm can produce multiple attitude images of ships during long-time observations, delivering clearer imagery even in complex motion scenarios.