基于矩阵填充理论的ISAR成像微多普勒抑制研究

发布时间：2018-03-12 07:09

本文选题：ISAR　切入点：微多普勒　出处：《哈尔滨工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：对逆合成孔径雷达(Inverse Synthetic Aperture Radar,ISAR)成像而言,目标上的微多普勒效应不仅会造成参数提取和运动误差补偿困难,而且会造成图像散焦,形成方位向干扰带,严重影响成像质量。因此,研究微多普勒抑制问题对ISAR的高精度成像有重要意义,同时在现代化战争的高精度目标识别对抗中,有重要应用价值。本文针对实际应用中的ISAR成像微多普勒抑制问题,建立合理有效的数学模型,并基于矩阵填充的理论,实现ISAR成像中微多普勒抑制研究。论文的主要内容及创新点如下:首先,确立了基于矩阵填充理论的微多普勒抑制研究方案。在详细分析了微多普勒抑制研究的应用背景和主要影响的基础上,结合微多普勒抑制研究的主要方法;分析得到单幅ISAR图像中,无论在实数域还是复数域微多普勒图像都具有低秩的特性,并且提出了基于低秩矩阵分解的单幅实数域ISAR像微多普勒抑制和复数域ISAR像微多普勒抑制的研究方案。此外,本文又从另一个角度去解决微多普勒抑制问题:首先生成一序列ISAR图像,将其重新排列成新的观测矩阵,根据差分阈值法进行低秩和稀疏的检测,并进行稀疏和低秩性分解,将目标和微多普勒进行分离,达到微多普勒抑制的效果。两类研究方案为全文的研究内容奠定了基础。其次,分析了基于转台模型RD成像振动和旋转微多普勒效应。针对早期“点目标”模型不足以反应目标微动信息的现状,以飞机转台为例,建立了ISAR成像的转台模型;结合经典RD成像算法,详细分析了振动和旋转两种典型微动造成的微多普勒效应;实现了微动散射点的回波分析和微多普勒效应干扰仿真,并给出了基于实测数据的微多普勒效应结果。再次,提出了基于低秩矩阵分解的ISAR微多普勒抑制方法。详细分析了飞机螺旋桨单散射点复数微多普勒图像和模值微多普勒图像的低秩性;针对ISAR图像稀疏和微多普勒图像低秩的特点,分析了矩阵的有效秩,并对含有微多普勒ISAR像进行SVD低秩拟合,保证微多普勒的低秩性;结合鲁棒性主分量分析(RPCA)理论和加速近邻梯度(APG)算法,实现了对ISAR像的模值RPCA和复数RPCA处理,并且在复数域中对APG算法进行了修正,重新构建了复数域的收缩算子。然后,分析对比了仿真中和实测数据中模值RPCA和复数RPCA处理效果。通过仿真和实测数据处理,验证了基于低秩矩阵分解方法实现ISAR微多普勒抑制的有效性;结合对图像信噪比和相干系数的分析,详细说明了本文所用方法对ISAR成像微多普勒效应的抑制效果良好。最后,在一定小转角内,连续成像一序列ISAR图像,保证目标基本没有转动,而微多普勒在这个期间内产生了大的变动。因此,将一序列图像中的每一幅图像转化成一个一维向量,重新组成新的观测矩阵,通过差分阈值法进行检测,得到目标为低秩矩阵,微多普勒为稀疏矩阵,并对其进行基于IALM算法和APG算法的低秩矩阵分解,并对实验结果进行分析对比。
[Abstract]:The inverse synthetic aperture radar (Inverse Synthetic Aperture Radar, ISAR) imaging, micro Doppler effect on the target will not only cause the parameter extraction and motion compensation is difficult, but also cause the defocus, form azimuth disturbance, seriously affect the image quality. Therefore, it has important significance to study the inhibition of high precision imaging for micro Doppler the problem of ISAR, and high precision of target recognition in the Modern War confrontation, has important application value. In this paper, ISAR imaging for the practical application of micro Doppler suppression, to establish a reasonable mathematical model based on matrix theory, the realization of ISAR imaging in micro Doppler suppression research. The main contents and innovation points are as follows: firstly, establish the matrix filling theory of micro Doppler suppression scheme based on research. In a detailed analysis of the micro Doppler inhibition studies should be used and the main background Based on the influence, combining with the main methods of micro Doppler inhibition; analysis of single ISAR image, has the characteristics of low rank in the real domain or the complex field of micro Doppler image, and put forward a real domain ISAR low rank matrix decomposition based on micro Doppler like inhibition and complex ISAR like micro Doppler suppression the research program. In addition, this paper also from another point of view to solve the problem of micro Doppler suppression: first generate a sequence of ISAR images, which can be rearranged into a new observation matrix, based on the detection of the differential threshold method of low rank and sparse, and sparse and low rank decomposition of the target and the micro Doppler to achieve separation of micro Doppler inhibitory effect. Two kinds of schemes for the study content of this thesis lays a foundation. Secondly, analyzes the vibration and rotation RD imaging turntable model based on micro Doppler effect for. Status quo of early "target" model do not reflect the target micro motion information, the aircraft simulator as example, established a model of ISAR imaging turntable; combined with the classic RD imaging algorithm, micro Doppler effect of vibration and rotation of two kinds of typical fretting caused by the detailed analysis of the implementation of the micro; scatter echo analysis and Simulation of micro Doppler effect the interference, and gives the micro Doppler effect results based on measured data. Thirdly, the low rank matrix decomposition ISAR suppression method based on micro Doppler. A detailed analysis of aircraft propeller single scattering point complex micro Doppler image and the modulus of low rank Doppler micro image; according to the characteristics of ISAR image sparse and micro Doppler image analysis of the effective rank, rank of matrix, and the micro Doppler ISAR like SVD containing low rank fitting, ensuring low rank of micro Doppler; combined with the analysis of robust principal component (RPCA) and The nearest neighbor and accelerating gradient (APG) algorithm, the realization of the ISAR like mode value RPCA and complex RPCA processing, which is modified and the APG algorithm in the complex domain, constructs contraction operator complex domain. Then, analysis and comparison of the RPCA and the complex number RPCA treatment effect simulation and measured data model value through simulation and experimental data processing, to verify the low rank matrix decomposition method to realize the effectiveness of ISAR suppression based on micro Doppler; combined with the analysis of the image signal-to-noise ratio and coherent coefficient, a detailed description of the method used in this paper the inhibitory effect on ISAR imaging and micro Doppler effect is good. Finally, in a small corner, continuous imaging of a sequence of ISAR images, which basically does not rotate, and the micro Doppler in this period had a big change. Therefore, will be transformed into a one-dimensional vector of each image in a sequence of images, again. The observation matrix is detected by differential threshold method, and the target is low rank matrix. Micro Doppler is sparse matrix. And the low rank matrix factorization based on IALM algorithm and APG algorithm is carried out, and the experimental results are analyzed and compared.

【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN957.52

【参考文献】