【作者】 付鹏；

【导师】 胡飞；

【作者基本信息】 华中科技大学 ， 电磁场与微波技术， 2023， 博士

【摘要】 综合孔径辐射计利用稀疏排列的小孔径天线阵列等效合成大孔径虚拟天线,可实现宽视场高分辨率成像,除被广泛地应用于射电天文与对地遥感外,安检成像等近场应用是综合孔径辐射计的另一个重要发展方向。现有的近场综合孔径辐射计成像算法延续远场假设将传播路径视为标量,基于傍轴近似研究近场成像问题。但在近场观测条件下,辐射源信号到达每个天线的传播方向均不一样,忽略该方向信息将带来近距离、宽视场观测条件下的离轴散焦问题以及倾斜因子引起的图像重建误差问题,难以高效地重建高质量的亮温图像。针对上述问题,论文将传播路径矢量化,基于干涉测量方法完善近场综合孔径可见度函数测量模型,建立近场综合孔径空间频率域表征模型与近场综合孔径卷积模型,提出基于角谱域变换的综合孔径辐射计近场成像方法。论文的主要工作概述如下:(1)为实现传播路径矢量化,论文基于标量衍射理论进一步完善Camps所提出的近场可见度函数,基于天线位置坐标描述近场可见度函数采样,将近场可见度函数测量模型的维度扩展为四维。理论分析表明,完善之后的近场可见度函数能更加全面地展现近场传播路径的矢量特征。(2)针对近场综合孔径成像中的离轴散焦问题,论文基于角谱理论对矢量化的传播路径进行分解,通过分析近场可见度函数的相位,建立近场综合孔径空间频率域表征模型。在此基础上,提出基于合成角谱的近场成像算法,通过获取传播方向信息实现近场成像。离轴点源成像实验证实了该算法可解决离轴散焦问题。(3)为减小近场综合孔径倾斜因子引起的图像重建误差,论文在合成角谱算法的基础之上,提出基于角谱域卷积的近场成像算法。通过建立包含倾斜因子的近场可见度函数与场景亮温分布的卷积关系模型,构造反演卷积核,将倾斜因子对可见度域的加权变为对角谱域的加权,提出基于角谱域卷积的近场成像算法。仿真实验验证了该算法可有效缓解倾斜因子引起的图像重建误差。(4)针对近场综合孔径快速成像应用需求,论文提出基于可见度-角谱域联合处理的近场快速成像算法。通过建立卷积核分解模型,基于十字阵、Y形阵以及旋转线阵等特殊阵列构型对近场可见度函数进行分类降维,提出基于可见度-角谱域联合处理的近场快速成像算法,基于卷积核分解模型分别在可见度域、角谱域实现相位补偿。实验研究表明,相比较于角谱域算法,所提算法在保障成像质量的同时,计算效率可提高两个数量级。综上所述,论文重点对近场条件下综合孔径辐射计成像的离轴散焦、倾斜因子加权等问题开展了研究,从完善近场可见度函数模型出发,基于角谱理论提出基于角谱域变换的综合孔径近场成像算法,并针对多种典型阵列提出快速成像算法,仿真和实验验证了所提算法的正确性和有效性。更多还原

【Abstract】 Synthetic aperture interferometric radiometer(SAIR)utilizes a sparsely arranged array of small aperture antennas to effectively simulate a large aperture virtual antenna,enabling wide field-of-view and high-resolution imaging.In addition to its widespread applications in radio astronomy and Earth remote sensing,near-field imaging,such as security check,represents another important development direction for the SAIR.Existing near-field SAIR imaging algorithms continue to treat the propagation path as a scalar and rely on paraxial approximation to study near-field imaging.However,under near-field observation conditions,the propagation direction of radiation signals reaching each antenna varies,and neglecting this directional information leads to the off-axis defocusing issue and the image reconstruction error issue caused by the oblique factor under close-range and wide field-ofview observations,making it challenging to efficiently reconstruct high-quality brightness temperature images.To address these problems,this dissertation vectorizes the propagation path,improves the measurement model of near-field SAIR visibility functions based on interferometry measurement methods,establishes the spatial frequency domain representation model and convolution model of near-field SAIR,and proposes a near-field imaging method based on angular spectrum domain transformation.The main contributions of this dissertation are summarized as follows:(1)To vectorize the propagation path,this dissertation further improves the near-field visibility function proposed by Camps based on scalar diffraction theory.The near-field visibility function sampling is described based on antenna coordinates,and the dimension of the measurement model of near-field visibility functions is expanded to four dimensions.Theoretical analysis shows that the improved near-field visibility function can more comprehensively represent the vector characteristics of near-field propagation paths.(2)To address the off-axis defocusing problem in near-field SAIR imaging,this dissertation decomposes the vectorized propagation path based on angular spectrum theory and establishes the spatial frequency domain representation model of near-field SAIR by analyzing the phase of the near-field visibility function.On this basis,a near-field imaging algorithm based on synthetic angular spectrum is proposed to achieve near-field imaging by acquiring directional information.Experimental verification of off-axis point source imaging confirms that the proposed algorithm can solve the off-axis defocusing problem.(3)To reduce image reconstruction errors caused by the oblique factor,this dissertation proposes a near-field imaging algorithm based on angular spectrum domain convolution on top of the synthetic angular spectrum algorithm.Firstly,the convolution relationship model between the near-field visibility function containing the oblique factor and the scene brightness temperature distribution is established.The inversion convolution kernel is constructed to convert the weighting of the oblique factor from the visibility domain to the angular spectrum domain,proposing the near-field imaging algorithm based on angular spectrum domain convolution.Simulation experiments verify that this algorithm can effectively alleviate image reconstruction errors caused by the inclination factor.(4)To meet the requirements of fast imaging in near-field SAIR,this dissertation proposes a near-field fast imaging algorithm based on visibility-angular spectrum domain joint processing.Firstly,the convolution kernel decomposition model is established,and the near-field visibility function is classified and reduced dimensionally based on special array configurations such as the cross array,Y-array,and rotated line array.The near-field fast imaging algorithm based on visibility-angular spectrum domain joint processing is proposed,and phase compensation is achieved separately in the visibility domain and the angular spectrum domain based on the convolution kernel decomposition model.Experimental studies show that compared to the angular spectrum domain algorithm,the proposed algorithm can improve computational efficiency by two orders of magnitude while ensuring imaging quality.In conclusion,this dissertation focuses on the off-axis defocusing and oblique factor weighting problems in near-field SAIR imaging.Starting from improving the near-field visibility function model,it proposes a near-field imaging algorithm based on angular spectrum domain transformation,and it also presents fast imaging algorithm for various typical arrays.The correctness and effectiveness of the proposed algorithms are verified through simulations and experiments.更多还原