Study On The 3D Imaging Technology For Terahertz Array Radar

Compared with two-dimensional(2D)radar imaging,three-dimensional(3D)target image with high resolution can provide more details for radar detection,which shows great value in many realms,including remote sensing and target recognition.Aiming at the requirements of Terahertz(THz)3D imaging in the applications of security inspection and military detection,this thesis focus on the 3D imaging technology for THz array radar,which dedicated to studing the imaging principles,imaging models and imaging algorithms.Moreover,the 3D azimuth-elevation imaging for THz radar,3D interferometry imaging using multiple antennas,the 3D imaging method based on rotational linear array,and circular array based 3D imaging method are proposed.The research work in this thesis can promote the development of THz radar imaging technology.Firstly,the azimuth-elevation imaging model in THz band is studied.The effects on the imaging performance,which caused by the nonlinear phenomenon of wideband signal in THz radar systems,are analyze,and then two nonlinear phase compensation methods are proposed.The azimuth-elevation 3D imaging model based on broadband signal is established,and the fast imaging method is proposed based on the non-uniform Fourier transform.Then,the imaging features of complex target models and rough surface model are analyzed at THz frequencies.For the first time,the azimuth-elevation imaging experiment was conducted to acheive the 3D imaging result of Siemens Star model,which validated the high resolution of THz radar.To eliminate the phase errors casued by nonlinearity signal and reduce the data amount of 3D sampling,the 3D surface imaging based on dual-frequency technique is derived,and the 3D imaging results of both rough cone and rough cube are achieved based on the electromagnetic calculation data.Subsequently,the 3D interferometric imaging using multiple antennas is proposed based on the interferometric processing techniques.The interferometric imaging model using two antennas is established and the interferometric imaging experiment in THz band is carried out,which contributes to the 3D imaging result of a scaled plane model.To take the advantage of wideband signal of THz radar,the joint processing method of dual-frequency band is proposed for L-shape baseline interferometry.The 3D imaging results of a scatter model of airplane are obtained based on the joint processing method,which demonstrate the improvement of interferometric accuracy.According to the imaging features of rough surface in THz band,the interferometric imaging modle is established based on multiple antennas,and the background removing method and the recursive phase unwrapping method are proposed.The 3D imaging performance are simulated and analyzed,and the interferometric imaging experiments are conducted,which achieve the reconstructed 3D image of a rough plate.Thirdly,to meet the requirements of real-time imaging in THz band,the 3D imaging method based on rotational linear array is developed.The 3D imaging model based on rotational linear array is established,and the imaging principles are derived from the perspective of spectral support.The 3D imaging performance are analyzed as a function of the signal bandwidth and the number of antennas.To achieve better imaging results,a fast two-dimensional imaging method in wavenumber domain and an optimization method of sparse array are proposed.Comprehensive simulations are carried to validate the feasibility and effectiveness of the imaging method and optimization method.Subsequently,the fast raw signal simulation method based on the inverse imaging processing is developed.Furthermore,the 3D imaging method in wavenumber domain is proposed to achieve the 3D imaging results of targets.Finally,with the introduction of array signal design techniques,a novel 3D imaging model based on circular array is proposed.The signal model of the circular array is designed,which leads to a differential radiation field in space,and then the 3D imaging resolution can be achieved with the combination of relative motion between radar and targets.The 3D imaging method based on convolution back-projection algorithm and power spectrum density estimation is proposed,which can improve the imaging accurancy and suppress noise.Moreover,the sparse representation model for circular array imaging is established,and the 3D reconstruction approach based on sparse Bayesian learning is proposed,which simplies the imaging implementation process.