Study On ISAR Imaging Of Multiple Maneuvering Targets

Inverse Synthetic Aperture Radar(ISAR)has the ability to realize detection and provide high-resolution images in all-time and all-weather conditions,and therefore widely applied in many fields.Unlike the imaging algorithms of Synthetic Aperture Radar(SAR),which concentrate researches on decoupling between the range and azimuth dimension,ISAR place more emphases on motion compensation techniques of non-cooperative target.When the target moves in steady-state,after motion compensation,high quality image can be conveniently obtained by utilizing the conventional Range-Doppler(RD)method.However,for target with complex maneuvering such as spin,nutation,procession and three-dimensional(3D)rotation etc.,the conventional method encounters obstacle.The radar often faces with multiple maneuvering targets,such as the aircraft formation,multiple-warhead ballistic missile,and the interference of unmanned aerial vehicle to civil aviation.Compared with the single target case,the different motions between multiple targets make the imaging process more complicated.Thus,study on ISAR imaging of multiple maneuvering targets has important significance and practical application value.More accurate image and higher resolution mean to provide more abundant information of target and scene.As the research moves along,there is an ever-growing demand for improving resolution to realize higher quality image output and interpretation.Therefore,it is necessary to research into the super-resolution technique without extending the bandwidth and antenna aperture.This dissertation revolves the above questions and the major works comprise the following aspects.Based on the analysis of the influence of target motion to radar signal,the correlation processing techniques in ISAR are summarized from three aspects,which are translational motion compensation,MTRC correction and imaging processing.Several common time-frequency analysis methods are presented,and the applications of them in non-stationary signal processing area are discussed.There are typically non-stationary characteristics in radar echoes of multiple targets,and each target exhibits different Doppler history.According to this feature,the ISAR imaging model is firstly established,and then the time-frequency analysis method is applied to get the images of multiple targets directly.ISAR imaging of multiple targets based on edge detection and Hough transform is discussed.Hough transform is employed to separate the range profiles of different targets,and then imaging processing is proceeded by approaches of single target.When faced with situation that range curvatures exist,this method becomes limited.Concerning this issue,the dissertation gives a new method to resolve it,which employs the generalized second-order Keystone transform as the basis.Through range walk and range curvature corrections,motion compensations of different targets are realized.Then the range profiles can be separated and images of multiple targets are obtained after azimuth compression.Besides,considering the possible Doppler ambiguity,the Doppler peaks detection method is introduced for ambiguity resolution.Multiple targets detection,high precision range acquisition,angle measurement are often carried out on the premise of accurate parameter estimation.When the changes of Doppler can be approximately considered as linear ones,signal analyses of multiple targets translate into parameter estimations of multi-component linear frequency modulated(LFM)signals.In consideration of higher-order motions of targets,the parameter estimation becomes complicated.This dissertation proposes a new algorithm for parameter estimation and ISAR imaging of multiple targets.The adjacent cross correlation function(ACCF)is firstly used to reduce the order of original signals.Then the generalized second-order Keystone transform and Fr FT are combined to estimate the motion parameters of different targets.From this,motion compensations can be accomplished.Finally,ISAR imaging of multiple targets is realized.Super-resolution technique based on the spatial spectrum estimation is researched.The scattering center extraction is implemented by two different ways,which are CLEAN and 2D U-ESPRIT.Performance comparisons among several methods are given.An algorithm for 3D super-resolution ISAR imaging of complex target is proposed,which is based on 2D U-ESPRIT.Firstly,the electromagnetic scattering characteristic and 3D echoes of the target are simulated.Secondly,the 2D U-ESPRIT is exploited to extract the locations of 3D scattering centers,with the amplitudes estimated via the least square method(LSM).Finally,3D super-resolution reconstruction of a complex target is achieved.High resolution range profile(HRRP)is reconstructed through the scattering centers,and is found to be in good agreement with the original range profile.By substituting the large numbers of echo data with several main scattering centers,the computation and memory costs are greatly reduced.Meanwhile,some useless information is filtered out,which can benefit the subsequent feature extraction,classification and identification significantly.When a ship target pushes on across the rough sea,it often shows complicated dynamical behaviors,such as 3D rotation(roll,pitch and yaw),which will generate different Doppler modulations at different times,and lead to typically non-stationary characteristics in radar echoes of the target.By analyzing the echo model and impact of 3D rotation,ISAR imaging of ship target with complex motion is studied via the time-frequency analysis method.Besides,some particular issues in ISAR imaging of multiple ship targets are discussed.