基于压缩感知的ISAR成像算法研究

发布时间：2018-03-27 16:14

本文选题：逆合成孔径雷达　切入点：压缩感知　出处：《哈尔滨工业大学》2014年硕士论文

【摘要】：逆合成孔径雷达(Inverse Synthetic Aperture Radar,ISAR)图像在目标识别、目标分类等多种应用中具有十分重要的地位,如何提高其距离和方位向的分辨率是重要的研究方向。目前,虽然有很多算法能达到高分辨率成像,但是当测量数据减少或者数据丢失时,这类算法已不再适用。然而,压缩感知(Compressive Sensing,CS)作为一种新兴的信号处理技术,可以有效地解决此问题。如何将压缩感知理论应用到雷达成像领域,进而获得高分辨率的雷达图像具有重要的研究意义。本文分析了三种稀疏重建算法,相比于其他算法,平滑零范数算法是速度最快,效果最好的重建算法,其复杂度可降低2-3个数量级。该算法主要的思想是用一种连续的函数去近似零范数,然后对函数进行分析求解。然而,平滑零范数的求解问题是非凸问题,其代价函数高度不平滑,具有较多的局部最小值。原始算法的初值采用的是最小二乘解,是一个较为粗略的解。针对此问题,本文结合加权思想,在第四章提出了一种改进的平滑零范数算法(MSL0),并通过大量仿真验证其性能。同时,分析了ISAR图像的稀疏特性,以信号重构为主要目的,将原始的成像分析模型转换成压缩感知成像模型,并有效地将提出的MSL0算法与ISAR成像结合起来,通过实测数据仿真验证其有效而稳健的性能。基于多测量矢量(Multiple Measurement Vector,MMV)方法及加权优化思想,本文提出了一种基于混合范数的稀疏重建算法(RWL2,0),大量仿真实验证明其具有较好的重建性能。围绕“如何对整个成像域进行处理”这个问题,建立了新的雷达回波信号分析模型,并将提出的RWL2,0算法应用到该模型。实验仿真证明,该成像算法具有较低的重构误差,较高的重构信噪比及分辨率较高的成像效果。
[Abstract]:Inverse Synthetic Aperture ISAR (inverse synthetic Aperture Radar) image plays an important role in many applications, such as target recognition, target classification and so on. How to improve the range and azimuth resolution is an important research direction. Although there are many algorithms that can achieve high resolution imaging, such algorithms are no longer applicable when measurement data is reduced or data is lost. However, compression sensing is a new signal processing technology. This problem can be solved effectively. How to apply the theory of compressed sensing to radar imaging and obtain high resolution radar images is of great significance. In this paper, three sparse reconstruction algorithms are analyzed, which are compared with other algorithms. The smoothing zero-norm algorithm is the fastest and most effective reconstruction algorithm, and its complexity can be reduced by 2-3 orders of magnitude. The main idea of the algorithm is to approximate the zero norm with a continuous function, and then analyze and solve the function. The problem of smoothing zero norm is a non-convex problem, whose cost function is not smooth and has more local minimum. The initial value of the original algorithm is the least square solution, which is a rough solution. In chapter 4, an improved zero-norm smoothing algorithm is proposed, and its performance is verified by a large number of simulations. At the same time, the sparse characteristics of ISAR images are analyzed, the main purpose of which is signal reconstruction. The original imaging analysis model is converted into a compression sensing imaging model, and the proposed MSL0 algorithm is effectively combined with the ISAR imaging. The effectiveness and robustness of the proposed method are verified by the simulation of measured data. Based on the multiple Measurement vector MMV method and the weighted optimization idea, the proposed method is used to improve the performance of the system. A sparse reconstruction algorithm based on mixed norm is presented in this paper. A large number of simulation experiments show that the algorithm has good reconstruction performance. A new radar echo signal analysis model is established around the problem of "how to deal with the whole imaging domain". The proposed RWL2O algorithm is applied to the model. The experimental results show that the proposed algorithm has lower reconstruction error, higher SNR and higher resolution.
【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】：2014
【分类号】：TN957.52

【相似文献】