Research On Target Detection And High Resolution Imaging Algorithms Of Terahertz Radar

The imaging resolution of terahertz radar is able to reach millimeter scale.By studying the specific target detection and imaging algorithms,the high imaging resolution characteristic of terahertz radar is made full use to acquire the high resolution range profile(HRRP),synthetic aperture radar(SAR)image,and inverse SAR(ISAR)image with more target feature information,which can effectively improve the capacity for automatic target recognition and is helpful to realize the practical application of terahertz radar.Therefore,it is worthy for studying the target detection and high resolution imaging techniques of terahertz radar.For more efficient target recognition of terahertz radar,it is the basic problem to accurately extract the HRRP of interested target and reconstruct the SAR,ISAR image with more target feature information from terahertz radar echoes.However,due to the terahertz radar characteristics of the high imaging resolution,the short wavelength and the influences of the complex detection environment,the aspect-dependent target,the traditional target detection and imaging algorithms are difficult to well satisfy the aforementioned requirements.Therefore,in this dissertation,after deeply studying the target detection and high resolution imaging techniques of terahertz radar,several nonparametric target detection methods for the range-spread target and high resolution imaging algorithms for the aspect-dependent target are proposed to accurately extract the HRRP of interested target and reconstruct the SAR,ISAR image with more target feature information from terahertz radar echoes.The main work and contributions are summarized as follows:1.In order to accurately extract the HRRP of interested target from terahertz radar echoes under the variable clutter environment,the estimation methods of the scatter information about the range-spread target and the nonparametric detection detectors for the range-spread target are investigated.After the maximum signal-to-clutter ratio principle estimates the range cells within the range-spread target and Otsu's method estimates dominant scatterers,the rank binary integration detector for the range-spread target based on the generalized sign detector(GSD)is proposed.However,the rank binary integration detector for the range-spread target based on GSD has a discrete false alarm probability,which is usually different from the set false alarm probability.Further,we deduce the rank binary integration detector for the range-spread target based on the weighted GSD,which has a continuous false alarm probability.Experiments show that the above detection methods can accurately extract the HRRP of interested target,within which the dominant scatterers are relatively densely distributed.2.When dominant scatterers are sparsely distributed and/or the signal-to-clutter ratio is lower,the maximum signal-to-clutter ratio principle can not accurately estimate the range cells within the range-spread target.To solve the above problem,the maximum rank and signal-to-clutter ratio principle is proposed.Combined the maximum rank and signal-to-clutter ratio principle with Otsu's method,the scatter information about the range-spread target can be more accurately estimated.Then,considering that the rank binary integration detector has a quantitative loss,the rank integration detector for the range-spread target based on GSD is proposed.While this detector is only applied to dominant sccatterers,the collapse loss can be avoided.Because the rank integration detector for the range-spread target based on GSD has a discrete false alarm probability,we deduce the rank integration detector for the range-spread target based on the weighted GSD.Experiments show that the above detection methods can accurately extract the HRRP of interested target.3.For high resolution imaging of the cooperative aspect-dependent target in the terahertz band,a wide-angle circular SAR algorithm based on the adaptive subaperture partition method is proposed.Gaussian function can be used to approximately represent the scattering energy distribution of aspect-dependent target versus aspect angle,which is tested by the experiments.Based on this approximately representation,the adaptive subaperture partition method is proposed.The composite circular SAR image is obtained with the noncoherent combination of these subaperture images on the basis of the generalized likelihood ratio test imaging approach.Experiments show that the proposed imaging algorithm can better keep the real reflectivity characteristics of target and the fusion image can provide more target feature information.4.For high resolution imaging of the non-cooperative aspect-dependent target in the terahertz band,a high resolution ISAR imaging method for the moving target with distributed terahertz radars is proposed.In the process of motion compensation,the envelope alignment method that a single pulse echo is chosen as the benchmark by the adjacent correlation coefficient to align envelope and the phase correction in patches method based on the weighted least-squares algorithm are proposed to make the image of the moving target well focused.After that,the ISAR imaging approach for the moving target with distributed terahertz radars is proposed to fuse these multi-angle images.Experiments show that the fusion image obtained by using the proposed algorithm can present the rough outline of the moving target and provide more target feature information.In summary,the relevant studies on target detection and high resolution imaging algorithms of terahertz radar are carried out in this dissertation,which provides a good foundation for automatic target recognition of terahertz radar and is of great significance for the practical application of terahertz radar.