Research On Moving Target Imaging Of FMCW SAR Mounted On Unmanned Aerial Vehicle

Two-dimensional high resolution image can be obtained by Synthetic aperture radar(SAR).Conventional SAR with pulse system has big volume and the cost is high.Frequency modulation continuous wave(FMCW)technology and SAR was combined.to overcome the disadvantages of conventional SAR,then FMCW SAR can be mounted on a small-size unmanned aerial vehicle(UAV).Nowadays,SAR has been employed to supervising system.In this system,moving target is more important than static object.Moving target imaging of FMCW SAR mounted on UAV is researched in this dissertation.FMCW SAR imaging,moving target detection,moving target imaging and moving target autofocus are analyzed specifically.The contents are divided into four parts shown as follows.At first,the echo signal of FMCW SAR is analyzed,and the Range-Doppler(RD)imaging algorithm and wavenumber domain imaging algorithm are analyzed specifically.The RD imaging algorithm is simple and efficient,but its accuracy is low.Wavenumber domain imaging algorithm is accurate since the signal model is accurate and approximation is not induced in its derivation.Extend Stolt interpolation is introduced to combine with conventional motion compensation algorithm.The azimuth dechirp is analyzed to improve the efficiency of extend Solt interpolation imaging algorithm in small aperture data.To get the same resolution with full aperture,the incoherent phase of different sub apertures is discussed,and sub images are added coherently after compensating the phase.The simulations are used to validate the effectiveness of relative algorithms,and practicality is illustrated by real data process.Secondly,echo signal of moving target is discussed.Moving target signal can not be extracted in Doppler domain.Further more,the signal of moving target and clutter in time-frequency domain is analyzed.And moving target and clutter are divided via their characteristics in time-frequency domain.Then moving target signal can be extracted.Real data process is used to validate the effectiveness and practicality of proposed algorithm.Finally,in order to get accurate moving target image,two dimensional frequency domain signal of moving target with uniform velocity is analyzed.The wavenumber domain translation is derived to focus the moving target.For motion parameter estimation,a motion parameter method based on entropy is proposed.And normalized relative speed(NRS)is induced to simplify the parameter searching.Further more,the signal of moving target with uniform acceleration is analyzed,and it can be focused by the proposed algorithm.Simulations are used to validate the effectiveness and accuracy of the algorithm,and real data process is used to illustrated its practicality.For moving target autofocus,phase gradient autofocus(PGA)is analyzed in detail.Combining the characteristic of focused moving target,the PGA can be applied to moving target autofocus,and the entropy in real data process explains the effectiveness.