Study On Imaging Algorithms For Airborne SAR With Complex Motion Trajectory

Synthetic aperture radar(SAR)equipped on the aircraft is known as airborne SAR.Airborne SAR is capable of working day and night under all weather conditions,long distance and high resolution imaging,which are the characteristics of SAR.Airborne SAR also has the advantages of lower budget,higher flexibility and shorter design cycle compared with SAR systems mounted on other platforms.After more than half a century of development,airborne SAR has been widely used in military and civil fields.The requirements for airborne SAR imaging performance are also constantly increasing in practical applications with the continuous development of microelectronics,radar signal processing and other technologies.At present the research on airborne SAR imaging mainly focuses on the linear motion trajectory and the corresponding theoretical research is also relatively mature in this condition.Nevertheless,the maneuverability requirements of the aircraft will be comparatively high in many practical situations.Under this condition the carrier plane is usually fly along relatively complex motion trajectories such as arcs,circles or irregular curves instead of a straight line.Ordinarily,it is difficult to achieve satisfactory imaging results or even unable to form a image using traditional frequency domain imaging algorithms conbined with corresponding motion compensation methods at this time.Therefore,this dissertation mainly carry out related research on fast time domain algorithms aims at airborne SAR imaging with complicated motion trajectory.The main research contents are:1.The second chapter firstly analyses the basic principle of the back projection algorithm(BPA),the earliest time domain algorithm.On this basis,the two dimensional spectrum distribution of the BPA imaging grid established in the Cartesian coordinate system is deduced and analyzed.And then compared with the two dimensional spectrum of the BPA imaging image in the polar coordinate system.Thus,fast back projection algorithm(FBPA)and fast factorized back projection algorithm(FFBPA)that are two fast time domain algorithms based on polar coordinate system are introduced.Then the relationship between the three algorithms of BPA,FBPA and FFBPA is explained.And the operations of the three are calculated.Finally,the three algorithms are further compared and analyzed through simulation experiments.The results of the simulation experiments demonstrate that the imaging rate of FFBPA is significantly better than that of FBPA and BPA.However,the actual rate increase is lower than the theoretical value due to the influence of two dimensional interpolation.In addition,it is also concluded that FFBPA actually loses a certain amount of imaging accuracy in exchange for an increase in imaging rate according to the simulation results.2.Two dimensional interpolation processing is required in the imaging process of the fast time domain algorithm based on the polar coordinate system,which lead to the decrease of imaging efficiency and imaging accuracy.As a consequence,a fast time domain algorithm based on the Cartesian coordinate system is introduced in the third chapter.First of all,the relationship between Cartesian grid distribution and two dimensional spectrum is analyzed in the third chapter.Then a spectrum compression method is introduced,which corrects the spectrum center shift in the time domain of range and azimuth firstly.The spectrum azimuth tilt is corrected in the range frequency domain and azimuth time domain thereby the space occupied by the spectrum is effectively reduced.When the spectrum compression method is embedded in the sub aperture fusion,the spectrum aliasing does not appear.So the fast time domain algorithm based on Cartesian coordinate system can imaging normally.Finally,the imaging performance and efficiency of the algorithm are analyzed by simulation experiments and compared with the fast time domain algorithm based on polar coordinate system.The simulation results indicate that the imaging efficiency and accuracy of the fast time domain algorithm based on Cartesian coordinate system are better than the algorithm based on polar coordinate system.3.On the basis of the fast time-domain algorithms studied in the previous two chapters,a motion compensation imaging method based on fast time domain algorithm is presented in the fourth chapter in view of the practical problems in airborne SAR imaging with complex motion trajectories.Firstly,how to establish the Fourier transform relationship between the image domain and the phase history domain in the fast time domain algorithm is analyzed.Thus,the feasibility of phase gradient autofocus(PGA)embedding fast time domain algorithm is verified.On this basis,the main steps of the PGA method are introduced and the process of combining the PGA method with the fast time-domain algorithm is analyzed.Then the motion compensation imaging process based on fast time domain algorithm is analyzed.Finally,the effectiveness of the method is verified by the raw data of X-band and C-band airborne circular SAR.The results show that this method can achieve well imaging effects.4.The fifth chapter firstly introduces the related hardware resources of the graphics processing unit(GPU)and the compute unified device architecture(CUDA)programming model with the NVDIA Quadro RTX 8000,which is the GPU carrying platform used in the experiments of this dissertation.On this basis,a GPU parallel implementation method of fast time domain algorithm is presented.The actual imaging efficiency of FFBPA based on GPU is proved by simulation experiments and the raw data of airborne circular SAR in the end.The results of simulation experiments and raw radar data processing indicate that the imaging efficiency of FFBPA has been significantly improved after using GPU acceleration.