Study On Scene And Moving Target Imaging For SAR Mounted On Maneuvering Platforms

Compared with optical imaging,Synthetic Aperture Radar(SAR)can perform imaging regardless of day and night,and has strong penetrating ability to bypass obstacles for imaging.The importance of SAR imaging has been shown in both military and civilian applications.Because the maneuvering platform can flexibly change the flight trajectory,so it is more suitable for the needs of military applications.The use of maneuvering platforms equipped with synthetic aperture radar has attracted more and more attention.Accordingly,the SAR imaging of maneuvering platforms has been widely studied.However,because the orbit of the maneuvering platform is generally different from that of the straight-line flight path,many imaging algorithms based on the linear model cannot be directly applied to the SAR imaging of the motorized platform.It is necessary to study the imaging algorithm of the curved orbit to meet the demand for high-precision and high-efficiency SAR imaging.Starting from the requirements of improving accuracy and efficiency,this paper analyzes the deficiencies of existing imaging algorithms and study the scene SAR imaging and moving target SAR imaging from the perspective of frequency and time domain processing.The main contents of the paper are summarized as follows:The descending flight of maneuvering platform can be generally divided into straight line flight and curved flight with acceleration.Aimed at the cross-range variation in the descending straight line flight,the range model of the target is converted into the inclined plane in Chapter 2,so the cross-range invariation is hold.Then the rotation of the coordinate axis is used to increase the utilization rate of frequency spectrum.Finally,by imaging in azimuth frequency domain,the possible scene aliasing in azimuth time domain imaging in short aperture time can be avoided.For the descending flight with acceleration,the range history with acceleration is first constructed,then the invariable term in cross-range direction and the phase error term caused by acceleration are separated.After compensating the error term,the conventional RMA imaging algorithm can be used for imaging.The simulation and measured data processing results verify the effectiveness of the proposed method.Compared with frequency domain algorithm,time domain algorithm can perform imaging in any orbit.With the rapid performance development of computing components,time domain algorithm is more suitable for the accurate SAR imaging of maneuvering platform.The calculation load of time-domain algorithm mainly comes from the calculation of the range history for all points,aimed at this problem,in the third chapter,by simplifying the calculation of the range history,time-domain fast back-projection(FBP)algorithm is accelarated.If the phase error caused by the slant range approximation does not exceed a certain value,the impact on the imaging quality is small.Therefore,using this principle,the scene is divided into many blocks,and the ranges of the points in each small block is appoximated by the plane wave hypothesis.By using this hypothesis,only the center point of the small scene needs to be accurately calculated.For other points in the same small scene,the range can be approximated by adding a constant.The results of simulation experiments show that this method can accelerate the time domain algorithm to some extent under the condition of ensuring accuracy.The time-domain BP algorithm can accurately image the scene under the non-linear track of the maneuvering platform.To increase the efficiency of BP algorithm,algorithms such as FBP and FFBP is proposed,which greatly improves the computational efficiency of time-domain algorithm.In the acceleration algorithms such as FBP and FFBP,the idea of grading is used to perform progressive coherent accumulation in the polar coordinate system.However,the projection between the local polar coordinate systems needs to be implemented by interpolation.The interpolation operation brings a large amount of computation,and it will introduce the accumulation of errors,resulting in the deterioration of imaging quality.To solve this problem,the width of the wavenumber spectrum is first analyzed.Then,the factor that make the wavenumber spectrum related to the scene width is obtained,therefore the compression kernel in the wavenumber spectrum is proposed.So the process of coherent accumulation can be oprated in the Cartesian coordinate.The images in the next stage are obtained by zero padding in the frequency domain and the addtion of low resolution images.By doing so,the amount of calculations introduced by interpolation is reduced and the accumulation errors are avoided.The effectiveness of the algorithm is verified by simulation and measured data processing results.In the process of SAR imaging for the maneuvering platform,besides obtaining the image of the scene,the defocus results of the moving targets may also be obtained.They need to be refocused to provide high-precision imaging results for later information extraction.Since the tracks of maneuvering platforms are curved in many cases,BP algorithm is required to compensate the influence of the orbital curve motion.Aimed at the accuration imaging of moving target,in Chapter 5,starting from the principle of the BP integration,the precise form of the moving target in the two-dimensional wavenumber domain is obtained.Then we directly transform the small block where the moving target is located into the two-dimensional wavenumber domain.In the domain,through the one-dimensional search of velocity information,the defocus phase in the two-dimensional wavenumber spectrum is compensated to obtain an accurate moving target imaging with the minimum entropy.Compared with the traditional algorithm,the complicated process of transforming the defocusing result of the moving target back to the echo domain for processing is avoided.The measured data and simulation results prove the effectiveness of the method.