Moving Target Detection For Multi-angle SAR Systems

Multi-angle Synthetic aperture radar(SAR)is a new type of SAR working mode to obtain target multi-dimensional information.By observing the interested area from different squint angles,multi-angle SAR has strong moving target detection ability gaining more target motion information,such as velocity,which makes it possible for parameter estimation.Thesis mainly studies the velocity estimation of moving target,clutter suppression and target trajectory extraction.The main contents are divided into following four parts:1.SAR moving target geometric model is established,and imaging process of moving target is analyzed,and azimuth and range offset of moving target is deduced,which lays a theoretical foundation for velocity estimation in multi-angle SAR system.2.Velocity estimation of moving target in multi-angle SAR system is discussed.Azimuth velocity estimation is completed by azimuth offset of the target on sequential images.According to different observation time interval,two different range velocity estimation algorithms are proposed,which achieve velocity estimation.Aiming at the factors such as positioning error and frequency modulation rate error,performance of the velocity estimation algorithm is analyzed,which not only proofs the effectiveness of the algorithm,but provides basis for the parameters design.3.Clutter suppression in multi-angle SAR system is studied.For different forms of systems,background difference method and Defocusing Shifted Difference(DSD)are given to suppress clutter.Aiming at trajectory extraction,the density based spatial clustering of applications with noise(DBSCAN)algorithm is utilized.The potential motion curve of target in multi-angle SAR image is analyzed,and a method based on trajectory discrimination is proposed to extract trajectory.The effectiveness of clutter suppression algorithm and trajectory extraction algorithm are verified by experiments.4.The system parameters design for multi-angle SAR is carried out.The main parameters include observation time interval,squint angle,and sub-aperture overlap,which are optimized to achieve high target detection performance.