Research On Motion Compensation And Imaging Algorithm Of Terahertz SAR Imaging

Since its introduction in the 1950s,synthetic aperture radar(SAR)has been rapidly developed because of its advantages of working under long-distance,all-weather and all-day environment,and the ability of obtaining two-dimensional high-resolution images of targets or scenes.It has been used in various fields such as military reconnaissance,environmental monitoring,and earthquake relief.SAR transmitts large time bandwidth product signals to obtain the high range resolution,and uses the synthetic aperture technique to obtain the high azimuth resolution.In recent years,with the rapid development of terahertz devices,terahertz band SAR has attracted attention because of its many advantages.Compared with the traditional microwave band SAR,the terahertz band SAR has the following advantages: it can easily transmit signals with larger time bandwidth product,thereby achieving higher range resolution;it can achieve the same resolution with a smaller synthetic aperture;the angular effect of the target under the illumination of the terahertz wave is more obvious;and the micro-Doppler in terahertz radar echo is more obvious,which is beneficial for detecting slow moving targets.Since the synthetic aperture is achieved by the radar platform motion,in order to obtain the ideal imaging result,the radar platform motion trajectory needs to be accurately recorded.Generally speaking,the radar motion parameters are recorded by the inertial guidance system in practice.When accurate radar motion trajectory can not be obtained from parameters provided by the inertial guidance system,motion compensation becomes the necessary content in SAR imaging signal processing.For example,the SAR platforms which uses aircraft such as helicopters,have high frequency motion(vibration)errors in addition to lower frequency motion errors due to airflow,engine vibration and other factors.The low-frequency motion error can be recorded by the inertial navigation system with good performance under certain conditions,and the high-frequency motion error is often difficult to record due to its high frequency and small amplitude.In the traditional X-band SAR imaging signal processing,since the wavelength is much larger than the high-frequency motion error,the influence of the high-frequency motion error on the phase of the SAR echo is negligible;however,in the terahertz-band SAR imaging signal processing,the vibration amplitude is equal or even bigger than the the transmited wavelength,which causes the significant influence on the echo phase.Therefore,in the motion compensation of terahertz SAR imaging,not only the influence of low-frequency motion error,but also the influence of high-frequency motion error must be considered.In this paper,we do a lot of work on the terahertz SAR imaging motion error compensation and imaging algorithm,which can be summarized as follows:1)The characteristics of terahertz SAR imaging are analyzed.Based on this,the motion error model of terahertz SAR imaging platform is established,and the form of echo signals which are affected by both high frequency vibration and low-frequency motion errors of the platform is derived.The terahertz SAR imaging process is performed by the Range-Doppler algorithm.After the range compression processing and range cell migration correction(RCMC),the echo signal of the range bin which the dominant scatterer locates in has the form of hybrid linear-sinusoidal frequency modulation signal.This causes ghosting phenomena after the azimuth processing,which seriously deteriorates the terahertz SAR imaging results.The influence of different parameters of the platform high-frequency vibration on the imaging results is analyzed and verified by simulation results.Finally,the microwave band SAR imaging motion compensation method,such as map drift method and phase gradient autofocus method,are introduced,and the limitations of using these methods directly in motion compensation for terahertz SAR imaging is analyzed.2)Study the high frequency vibrationi compensation for terahertz SAR imaging based on single-component harmonic vibration model.When the platform vibration consists of single component harmonic vibration,after the doppler frequency rate term is removed by constructing the reference signal according to the known parameters,the echo signal of the range bin which the dominant scatterer locates in has the form of single component sinusoidal frequency modulation(SFM)signal.Based on this conclusion,the parameter estimation algorithms for SFM signals are studied.Firstly,based on the existing inverse Radon transform-based method,we add the step of extracting the time-frequency ridge of the time-frequency representation to reduce the computating load.Secondly,the parameter estimation of SFM signal based on adaptive Chirplet decomposition is proposed.Thirdly,the discrete sinusoidal frequency modulation transform and the global optimization algorithms,such ase the simulated annealing algorithm and the particle swarm optimization algorithm,are combined to achieve the parameter estimation of the SFM signal with a low computational complexity.Aiming at the parameter estimation of hybrid linear-sinusoidal frequency modulation signals,a new transform,namely the extended Radon transform is proposed,and its application in parameter estimation for SFM signal is invesitgated.At last,contrast and the application range of these methods are given.3)Study the high frequency vibration compensation for terahertz SAR imaging based on multi-component harmonic vibration model.When the platform vibration consists of multi-component harmonic vibration,after the doppler frequency rate term is removed by constructing the reference signal according to the known parameters,the echo signal of the range bin which the dominant scatterer locates in has the form of multiple sinusoidal phase modulation signal.The parameter estimation methods of this kinds of signals are studied.Three different parameter estimation algorithms are proposed for this form of signal.Firstly,the improved discrete sinusoidal frequency modulation method is combined with the optimization algorithm such as simulated annealing algorithm to estimate the parameters of the signal under the premise of low computational complexity.Secondly,the combination of sinusoidal frequency modulation Fourier transform and maximum likelihood is used to realize the parameter estimation of the signal;the third is to combine the short-time Fourier transform and the maximum likelihood method to estimate the parameters of the signal;finally,the similarities,differences and application range of the three methods are analyzed.4)Study the terahertz SAR motion compensation and imaging algorithm based on indoor measured data.The proposed motion compensation algorithm considers both the low-frequency motion error and the high-frequency motion error compensation.The algorithm firstly uses the parameter estimation algorithm mentioned above to complete the parameter estimation of the platform high-frequency vibration error,and then constructs the reference signal to compensate for the platform high frequency vibration error.The strip phase gradient autofocus algorithm is used to compensate the low frequency motion error of the platform.Finally,the effectiveness of this new algorithm is verified by the indoor measured data.