An improved two-step motion compensation method based on raw data

Motion compensation (MOCO) is the key step in airborne Synthetic Aperture Radar (SAR) imaging processing. Motion errors can be obtained from the navigation data which is acquired from the inertial navigation system (INS) mounted on the plane. However, the accuracy of the navigation data is limited which is far from application requirement for high resolution system. Accordingly, raw-data based MOCO approach is necessary in airborne SAR azimuth focusing which is defined as autofocus method. The autofocus method includes no-parametric techniques such as phase gradient autofocus (PGA) method and parametric techniques including contrast optimization algorithm (COA), map drift algorithm (MDA), etc. Based on the conventional two-step MOCO method, this paper proposes an improved two-step MOCO technique combining the parametric and the no-parametric techniques to alleviate the dependence on the navigation measurement. Firstly, an improved sub-aperture COA is implemented to acquire the Doppler rate accurately which can be utilized in the estimation of motion errors. In the next step, PGA is applied to eliminate the residual phase errors. Imaging results on real strip-mode airborne SAR data validate the proposed MOCO approach.