逆合成孔径雷达二维及三维成像方法研究

发布时间：2018-06-27 22:32

本文选题：逆合成孔径雷达 + 相位误差补偿　；参考：《西安电子科技大学》2016年博士论文

【摘要】：逆合成孔径雷达(Inverse Synthetic Aperture Radar,ISAR)成像可获得目标的一维、二维和三维高分辨率成像结果,提供观测目标的尺寸和结构等丰富的特征信息,它在空间态势感知和防空反导等领域中发挥着重要的作用。经过六十多年的长足发展,ISAR成像的基本理论较为成熟。但是,随着ISAR系统的不断发展和目标观测需求的不断提高,ISAR二维和三维成像尚有不少问题亟待解决。首先,随着当前ISAR信号频率向高频段方向的不断发展,雷达对发射及接收系统的相位噪声和非线性影响更为敏感,将造成回波的距离维相位误差,进而降低目标的ISAR成像质量。其次,对于远作用距离和弱后向散射系数的目标进行观测时,其回波的信噪比较低,传统的ISAR成像方法难以完成精确的平动补偿与成像。再次,对编队飞行目标的复杂观测场景进行成像时,子目标相对雷达视线的运动形式存在一定的差别,而其回波严重耦合,如何实现编队飞行目标良好的聚焦处理是目前雷达成像研究领域中的难题。同时,序列ISAR图像包含了更多目标的信息,如何基于序列ISAR图像实现目标的精确定标和三维重构,是目前ISAR雷达成像研究领域中的挑战性问题。上述这些问题不仅是雷达成像研究领域的难题,同时也是制约实际ISAR系统发展与应用的技术瓶颈。因此,本文研究的ISAR二维与三维成像方法具有重要的理论意义和应用价值。本文在国家自然科学基金、国家863课题与横向课题等项目的资助下,结合实际ISAR系统的发展趋势和复杂目标观测的迫切需求,针对ISAR距离维相位误差补偿、低信噪比下ISAR成像、多目标ISAR成像、图像定标和目标三维重构等问题进行了深入的研究,取得一些理论研究成果,并获得聚焦良好的仿真与实测数据处理结果,进而提升了ISAR的信息获取能力。本文的具体内容可分为以下六个部分：1.首先,对ISAR成像的基本原理进行介绍,构建了ISAR成像几何模型和信号模型。然后,对ISAR成像处理过程中平动补偿的物理意义和常用方法进行了详细的讨论,阐述了越距离单元徙动(Migration Through Range Cells,MTRC)现象的产生原因、影响以及相应的校正方法,为后文的二维与三维ISAR成像方法研究打下理论基础。2.针对高波段宽带ISAR的距离维非线性相位误差严重影响目标成像质量的问题,提出了一种精确的基于最小熵与自聚焦处理的距离维相位误差快速补偿方法,获得高精度的目标距离维聚焦处理结果。首先,建立了ISAR回波距离维相位误差的高阶多项式模型,详细分析了不同阶次相位误差对距离压缩结果聚焦度的影响。然后,将距离压缩结果的最小熵作为目标函数,利用启发式搜索算法,对二阶距离维相位误差进行准确高效的估计和补偿。最后,采用距离维自聚焦处理以准确估计和校正三阶及三阶以上的高次距离维相位误差,以获得高精度的距离维相位误差补偿结果。仿真数据处理结果验证了距离维相位误差对ISAR目标成像聚焦度的影响和本章所提出方法的有效性。3.针对低信噪比下传统ISAR成像方法难以精确平动补偿与成像的难题,提出了一种基于频谱信号相位求导和粒子群优化的自适应的低信噪比下ISAR平动补偿与成像方法,获得了低信噪比情况下聚焦良好的ISAR实测数据成像结果。首先,结合相干的ISAR原始雷达回波,建立了目标平动分量的高阶多项式模型,可有效对不同的目标平动形式进行表示。接着,基于距离维回波频谱相位求导提出了一种目标平动分量多项式系数的粗估计方法,并将其结果作为后续利用粒子群优化进行平动分量精确估计的参考初始值,以提高运算效率。然后,利用粒子群优化(Particle Swarm Optimization, PSO)算法对目标平动分量进行精确估计,该算法实现简便,有计算效率和估计精度高等优点。最后,分析成像结果聚焦度随目标平动分量重构精度的变化特性,以自适应地确定目标平动分量多项式模型的阶数,因此,本章所提出方法的通用性更强。仿真和实测数据处理结果验证了本章所提出方法的有效性。4.针对编队飞行目标等复杂观测场景中不同子目标回波严重耦合的难题,提出了基于粒子群优化和改进CLEAN算法的多目标ISAR成像方法。首先,针对传统多目标ISAR成像算法将不同目标的运动限定为基本一致或者完全不同的局限性,将多目标划分为多个组目标,同一组目标内不同子目标的运动形式和参数是一致的,该模型的通用性更强。其次,构建组目标平动分量的多项式模型,迭代利用PSO算法对不同组目标的平动分量进行准确的估计和补偿,得到不同组目标的粗聚焦成像结果。然后,提出了改进CLEAN算法对该组目标成像结果进行提取,在保留被提取组目标图像完整性的同时,对虚假点和噪声进行了更有效的压制。最后,利用聚类数估计和K-均值算法对同一组目标内的子目标进行分割和提取,进而通过单目标成像算法逐个得到每个子目标的良好聚焦成像结果。仿真实验结果验证了本章所提出方法的有效性。5.针对ISAR成像处理的横向定标难题,基于目标等效旋转角速度与ISAR散射点回波二次相位系数的解析关系,提出了一种基于离散多项式相位变换的ISAR图像定标方法,获得高精度的Yak42飞机实测数据定标处理结果。首先,对距离单元回波进行了深入的分析,详细推导了目标等效转动角速度与散射点二次相位系数的解析关系。其次,计算不同距离单元回波的归一化幅度方差,并选取值较小的距离单元。接着,利用频域加窗技术对上述距离单元中的强散射点频谱进行提取,同时给出了窗长的自适应确定方法。之后,提出采用二阶离散多项式相位变换(Discrete Polynomial-Phase Transform, DPT)对强散射点回波中的二次相位系数进行估计,相比传统基于最大对比度搜索和尺度Radon-Winger变换(Radon-Wigner Transform, RWT)的方法,该方法精度和计算效率较高,且实现简单。最后,利用最小均方误差(Least Square Error,LSE)估计得到目标的等效转动角速度,实现ISAR图像精确的方位定标。仿真和实测数据的定标结果验证了本章所提出方法的有效性。6.针对现有的基于序列ISAR图像的目标三维重构算法存在尺度模糊和无法定标的问题,提出了一种基于松弛约束分解的目标三维重构和图像定标联合处理方法,其性能优于现有的基于因式分解的三维重构方法。首先,构建了ISAR目标的三维运动和成像模型,并详细推导了目标三维结构向成像平面的投影方程解析表达。之后,将大转角的ISAR回波分段处理,以降低后续成像处理的难度,得到未定标的序列ISAR图像,进而得到目标散射点的航迹矩阵。然后,提出了改进的因式分解方法,并利用其对散射点航迹矩阵进行分解,该方法有效避免了航迹矩阵中散射点未定标的方位位置对三维重构精度的影响。接着,基于重构的投影向量,通过最小均方误差估计方法对目标等效转动角速度进行估计,并重新对散射点的航迹矩阵进行方位尺度校正。最后,迭代利用松弛约束的因式分解方法和目标等效转动角速度估计方法来提高算法精度,进而联合实现目标三维结构的重构和ISAR图像的定标。仿真数据处理结果表明了本章所提出方法的性能优于现有的基于因式分解的三维重构方法。
[Abstract]:Inverse Synthetic Aperture Radar (ISAR) imaging can obtain one dimensional, two-dimensional and three-dimensional high-resolution imaging results of the target, providing rich feature information such as the size and structure of the observation target. It plays an important role in space situational awareness and air defense and antimissile. After more than 60 years of rapid development, The basic theory of ISAR imaging is more mature. However, with the continuous development of the ISAR system and the continuous improvement of the target observation demand, there are still many problems to be solved in ISAR and 3D imaging. First, with the continuous development of the current ISAR signal frequency to the high frequency section, the phase noise and nonlinearity of the emitter and receiving system The effect is more sensitive, which will cause the distance dimension phase error of the echo, and then reduce the ISAR imaging quality of the target. Secondly, when the target of the distance and the backward scattering coefficient is observed, the signal to noise of the echo is low, and the traditional ISAR imaging method is difficult to complete the accurate translational compensation and imaging. Again, the formation flight target is made. When the complex observation scene is imaging, there is a certain difference in the motion form of the sub target relative to the radar line of sight, and the echo is seriously coupled. How to realize the good focus processing of the formation flight target is a difficult problem in the field of radar imaging research. At the same time, the sequence ISAR image contains more information of the target, and how to base on the sequence ISAR It is a challenging problem in the field of ISAR radar imaging to realize the accurate target calibration and 3D reconstruction of the target. These problems are not only a difficult problem in the field of radar imaging research, but also a technical bottleneck restricting the development and application of the actual ISAR system. Therefore, the two dimensional and three-dimensional imaging methods of this paper are important for the research of this paper. Under the support of the National Natural Science Foundation, the National 863 subject and the horizontal project, this paper combines the development trend of the actual ISAR system and the urgent needs of the complex target observation, aiming at the ISAR distance dimension phase error compensation, the low signal to noise ratio ISAR imaging, the multi target ISAR imaging, the image calibration and the target. Some theoretical research results are carried out, some theoretical research results are obtained, and the results of good simulation and actual data processing are obtained. The information acquisition ability of ISAR can be improved. The specific content of this paper can be divided into six parts: 1. first, the basic principle of ISAR imaging is introduced, and the geometry of ISAR imaging geometry is constructed. Then, the physical meaning and common methods of the translational compensation in the ISAR imaging process are discussed in detail. The causes of the Migration Through Range Cells (MTRC) phenomenon, the influence and the corresponding correction method are expounded, and the two-dimensional and three-dimensional ISAR imaging methods of the later text are studied. In view of the problem that the target imaging quality is seriously affected by the distance dimension nonlinear phase error of the high wave band wideband ISAR, a fast compensation method of distance dimension phase error based on the minimum entropy and autofocus processing is proposed, and the high precision target distance dimension focusing processing results are obtained. First, the ISAR echo distance is established. In the high order polynomial model of phase error, the influence of different order phase errors on the focus of distance compression is analyzed in detail. Then, the minimum entropy of the distance compression result is used as the objective function, and the two order distance dimension phase error is accurately and efficiently estimated and compensated by the heuristic search algorithm. Finally, the distance dimension is used. The self focusing process is used to accurately estimate and correct the high order distance dimension phase error of the three order and above three order to obtain the high precision distance dimension phase error compensation results. The simulation data processing results verify the effect of the distance dimension phase error on the imaging focusing degree of the ISAR target and the validity of the method proposed in this chapter under the low signal to noise ratio. The traditional ISAR imaging method is difficult for accurate translation compensation and imaging. A ISAR translational compensation and imaging method based on the phase guidance of the spectrum signal and particle swarm optimization is proposed in the adaptive low signal to noise ratio (SNR). The imaging results of the ISAR measured data with good focus in the case of low signal to noise ratio are obtained. First, the coherent ISAR original mine is combined. The high order polynomial model of the target translational component is established, and the different target translational forms can be effectively expressed. Then, a rough estimation method for the polynomial coefficients of the target translational component is proposed based on the spectral phase derivation of the range dimension echo spectrum, and the result is made for the subsequent use of particle swarm optimization to make the translation component accurate. The estimated reference initial value is used to improve the operational efficiency. Then, the Particle Swarm Optimization (PSO) algorithm is used to accurately estimate the target translational components. The algorithm is simple to implement, has the advantages of high computational efficiency and high estimation precision. Finally, the changes of the imaging results focusing degree vary with the accuracy of the target translation component. In order to determine the order of the polynomial model of the target translation component adaptively, the method proposed in this chapter is more versatile. The simulation and measured data processing results verify that the effectiveness of the method proposed in this chapter.4. is based on the difficult problem of the serious coupling of the different sub targets in the complex observation scenes such as the formation flying targets. The multi-objective ISAR imaging method of particle swarm optimization and CLEAN algorithm is improved. Firstly, the traditional multi-objective ISAR imaging algorithm defines the motion of different targets as basically consistent or completely different limitations, and divides the multi target into multiple group targets, and the motion forms and parameters of the different subtargets in the same target are the same, the model is the same. Secondly, the polynomial model of the target translation component of the group is constructed, and the PSO algorithm is used to accurately estimate and compensate the translational components of different groups of targets, and the results of the rough focusing imaging of different groups of targets are obtained. Then, the improved CLEAN algorithm is proposed to extract the target imaging results of the group and be extracted and extracted. With the integrity of the target image, the false points and noise are more effectively suppressed. Finally, the clustering number estimation and K- mean algorithm are used to segment and extract the sub targets in the same set of targets, and then the good focusing imaging results of each sub target are obtained by single target imaging algorithm. The simulation results verify the results. The validity of the method proposed in this chapter.5. is based on the analytic relationship between the two phase coefficients of the target equivalent rotation angular velocity and the ISAR scattering point echo on the basis of the analytic relationship between the two phase coefficients of the target equivalent rotation angular velocity and the scattering point echo of the ISAR. A ISAR image calibration method based on the discrete polynomial phase transformation is proposed to obtain the high precision calibration of the measured data of the Yak42 aircraft. First, the echo of the distance unit is deeply analyzed. The analytic relation between the target equivalent rotation angular velocity and the two phase coefficient of the scattering point is derived in detail. Secondly, the normalized amplitude variance of the echo of different distance units is calculated and the distance unit with a smaller value is selected. Then, the range unit is used in the frequency domain to add the window technique to the distance unit. The strong scattering point spectrum is extracted and an adaptive method for determining the length of the window is given. After that, the two order discrete polynomial phase transformation (Discrete Polynomial-Phase Transform, DPT) is used to estimate the two phase coefficients of the strong scattering point echo, compared with the traditional maximum contrast search and the scale Radon-Winger transformation (Rad). The method of on-Wigner Transform, RWT) has high accuracy and efficiency, and it is simple to realize. Finally, the equivalent rotation angular velocity of the target is obtained by using the minimum mean square error (Least Square Error, LSE) to realize the accurate azimuth calibration of the ISAR image. The simulation and the calibration results of the real data verify the effectiveness of the method proposed in this chapter. In view of the problem that the existing target 3D reconstruction algorithm based on the sequence ISAR image exists the problem of scale fuzzy and uncalibrated, a method of joint processing of target 3D reconstruction and image calibration based on relaxation constraint decomposition is proposed. The performance of.6. is superior to the existing 3D reconstruction method based on factorization. First, the ISAR target is constructed. The three-dimensional motion and imaging model of the target are described in detail, and the projection equation of the three-dimensional structure of the target to the imaging plane is derived in detail. After that, the ISAR echo of the large rotation angle is segmented to reduce the difficulty of the subsequent imaging processing, and the unfixed sequence ISAR image is obtained, and then the track matrix of the target scattering point is obtained. Then, the improved cause is proposed. The method of decomposition is used to decompose the scattering point track matrix. This method effectively avoids the influence of the azimuth position of the scattering point in the track matrix on the three-dimensional reconstruction accuracy. Then, based on the reconstructed projection vector, the minimum mean square error estimation method is used to estimate the equivalent rotational angular velocity of the target, and the dispersion is rearranged. The path matrix of the point is corrected for azimuth scale. Finally, the algorithm is iteratively used to improve the accuracy of the algorithm and the target equivalent rotation angular velocity estimation method, and then the reconstruction of the three-dimensional structure of the target and the calibration of the ISAR image are combined. The results of the simulation data processing show that the performance of the proposed method is superior to that of this chapter. The existing three-dimensional reconstruction method based on factorization.
【学位授予单位】：西安电子科技大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TN958

【相似文献】