Study On Refocusing Methods For Ship Target From Spaceborne Radar

Spaceborne radar gains a wide attention,which is attributed to the advantages such as wide-field detection,high resolution at long range,and strong anti-attack capability.When the radar is imaging the marine target with complex non-linear motion,which is blurred by ocean waves,the following problems will happened.Firstly,when the ship motion model is established,not only the ship translation motion but the six DOF motion have to be concerned.Secondly,the common motion compensation algorithms are unable to refocus the SAR images,when the non-linear motion leads to non-stationary echo.Thirdly,in order to gain the high azimuth resolution,the synthetic aperture time is required to extend.But the longer integration time,the more complex non-linear motion is brought in.Therefore,it is necessary to study the ship targets’ refocusing algorithm,which has significant meaning for imaging marine targets.This thesis is focusing on not only the defocused SAR images,which is caused by the non-linear motion,but the estimation of ship target’s motion parameter or compensation of the phase error.The main work is highlighted as follows:In the first,the thesis combing the marine targets six DOF motion characteristics with the model of orbit and the spaceborne radar observation model to establish the precise echo model.Then the thesis analysis the factors which lead to a blurred ship image,and identify the applicability of traditional ISAR compensation methods.Secondly,aiming at the echo with high signal to noise ratio(SNR),the thesis study a new time frequency curve extracted method based on compensating the translation motion.This method combines the vertical synchrosqueezing transform with Bernoulli filter.At first step,the method synchro-squeezes the time frequency area signal data to concentrate the time frequency curves and increase the accurate degree of instantaneous frequency.And then establish the state translation model on the basis of the time frequency curve.Using the Bernoulli filter based on random finite sets to achieve the high precise estimation of the scatters’ micro-motion parameters.The study of theory and the results of simulation prove that this method can avoid the low accuracy problem and the matter of multi-curve tracking errors.Compared with inverse radon transform and particle filter method,this algorithm has advantages of higher precision,lower computation and stronger timeliness.Finally,for the low SNR signal,the thesis compare the searching matching filter method in full-dimensional space without compensating envelope with general radon fourier transform joint searching parameter method.Based on that,the thesis put forward a method using both the motion characteristics of rigid target and the image entropy minimum criterion for optimization constraints.The method concerns about differences between ship target and air plane,then establishes an accurate echo model.A joint salient points rough estimation method is studied,which is data-driven,in order to speed up the coordinate rotation integration speed.The study of theory and the results of simulation prove that ship target posture can be indicated by the integration results.And this method has a lot of superiorities such as low computation,high precision,well compensation of blurred ship images and strong noise robustness.