Research On Moving Target Imaging Of Non-uniform Velocity UAV SAR

Unmanned aerial vehicle synthetic aperture radar(UAV SAR)is a high-resolution imaging system that carries SAR on UAV platform.Because of its high cost performance,the use does not need to consider the pilot factor,simple daily maintenance,and has the characteristics of light weight,small volume,mobility and flexibility,it has a wide application prospect.However,in practical application,due to its light weight,UAV is easy to be affected by low-level atmospheric turbulence,which usually deviates from the nominal constant speed linear flight trajectory.Due to the limitation of manufacturing cost,UAV SAR usually adopts low pulse repetition rate,which reduces the amount of data,but usually can not meet the requirements of Nyquist sampling theorem,resulting in the problem of undersampling.Because UAV SAR is affected by flight trajectory deviation and undersampling,the traditional moving target detection algorithm and moving target imaging algorithm based on uniform motion model will no longer be applicable.In this thesis,the moving target parameter estimation and imaging of UAV SAR under non-uniform linear flight trajectory and undersampling are studied,as follows:Aiming at the parameter estimation problem of UAV SAR under non-uniform linear flight and undersampling,this thesis studies a moving target parameter estimation method based on compressed sensing(CS)under non-uniform motion.This method constructs a signal model with moving target parameter dictionary by using the relationship between azimuth undersampling echo signal and moving target parameters under UAV SAR nonuniform trajectory,Based on the model,the optimization function of moving target parameter estimation is generated,so that the parameter estimation problem is transformed into an optimization problem with norm constraints,and the parameter estimation of moving target is realized by using approximate gradient descent(PGD)algorithm.In addition,this thesis also studies the traditional moving target parameter estimation algorithm based on scaled Fourier transform(SCFT),and compares the computational complexity of the above two parameter estimation algorithms.The superiority of moving target parameter estimation algorithm based on CS is verified by simulation and measured data.In order to achieve high-efficiency and high-quality moving target imaging under azimuth undersampling and non-uniform motion of UAV SAR platform,a moving target imaging algorithm based on fast iterative shrinkage threshold algorithm(FISTA)and nonuniform fast Fourier transform(NUFFT)is studied in this thesis.Using the accurate moving target frequency domain signal model,the moving target imaging problem is transformed into an optimization problem constrained by Frobenius norm and norm,and then the optimization problem is solved quickly by FISTA and NUFFT.The computational complexity of the algorithm is analyzed,and the effectiveness of the imaging algorithm for moving target imaging on UAV SAR platform is verified by simulation and measured data.