基于一维距离像的舰船目标识别技术研究

发布时间：2018-02-13 04:34

本文关键词： 舰船目标识别一维距离像 K分布海杂波海杂波抑制　出处：《电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着高分辨雷达技术的日趋成熟,雷达目标自动识别技术在对海作战、港口管理和海上搜救等方面发挥了重要的作用。高分辨一维距离像能够反映目标散射中心沿雷达径向的分布情况,具备数据容易获取、处理相对简单等优点,受到了国内外许多学者的关注和研究。基于一维距离像的目标识别算法,大多集中于飞机、导弹等空中目标,而舰船一维距离像受到的干扰主要来自海杂波,其特性与低分辨雷达瑞利杂波有所不同,大量研究表明,高分辨雷达海杂波的分布特性可以用K分布模型准确描述,因此,本文对K分布海杂波背景下的舰船目标一维距离像识别进行了研究,主要工作内容如下:1.以雷达目标的点散射中心模型为基础,论述了解线性调频技术的基本原理,并对该技术的处理过程进行了计算机仿真,得到了三类舰船目标在宽带雷达下的一维距离像。研究分析了舰船目标一维距离像的三大敏感特性及其克服方法,并进行了仿真验证。2.介绍了海杂波的K分布模型,研究了其对舰船目标一维距离像的成像以及形状参数和尺度参数对特征提取的影响,并和同等信杂比条件下的瑞利杂波进行了对比分析。3.论述了模式识别领域中几种经典的特征提取和分类器设计方法,结合仿真数据,分析了这些方法各自的优缺点。利用不同的特征提取方法和分类器,对K分布海杂波背景下的舰船目标进行一维距离像识别,通过蒙特卡洛实验,对比了上述特征提取方法和分类器的性能及计算效率。4.研究了基于AR模型和基于奇异值分解的两种海杂波抑制技术,并将两者应用于低信杂比情况下的舰船目标一维距离像识别,通过理论分析和仿真实验,对比了两种海杂波抑制技术各自的优缺点,其中基于AR模型的海杂波抑制技术的性能相对更好。舰船目标一维距离像受海杂波影响,导致特征的类间趋同性增加,通过引入海杂波抑制技术,可以提高低信杂比条件下的舰船目标一维距离像识别性能。
[Abstract]:With the development of high-resolution radar technology, automatic radar target recognition technology is fighting against the sea. Port management and maritime search and rescue play an important role. High-resolution one-dimensional range profiles can reflect the radial distribution of the target scattering center along the radar, and have the advantages of easy data acquisition and relatively simple processing. Many scholars at home and abroad have paid close attention to it. The target recognition algorithm based on one-dimensional range profile is mostly focused on aircraft, missile and other air targets, while the interference of ship one-dimensional range image is mainly from sea clutter. It is different from Rayleigh clutter in low resolution radar. A large number of studies show that the distribution characteristics of sea clutter in high resolution radar can be accurately described by K distribution model. In this paper, the recognition of one-dimensional range profile of ship target in the background of K-distributed sea clutter is studied. The main work is as follows: 1. Based on the point scattering center model of radar target, the basic principle of linear frequency modulation (LFM) technology is discussed. The processing process of this technology is simulated, and the one-dimensional range profile of three kinds of ship targets under wideband radar is obtained. The three sensitive characteristics of the one-dimensional range profile of ship targets and their overcoming methods are studied and analyzed. Secondly, the K distribution model of sea clutter is introduced, and the influence of shape parameters and scale parameters on feature extraction is studied. And compared with Rayleigh clutter under the same signal-to-clutter ratio. 3. Several classical feature extraction and classifier design methods in the field of pattern recognition are discussed, and the simulation data are combined. The advantages and disadvantages of these methods are analyzed. Different feature extraction methods and classifiers are used to identify ship targets under the background of K-distributed sea clutter. The performance and computational efficiency of the above feature extraction methods and classifiers are compared. 4. Two kinds of sea clutter suppression techniques based on AR model and singular value decomposition are studied. The two methods are applied to the range profile recognition of ship targets with low signal-to-clutter ratio. Through theoretical analysis and simulation experiments, the advantages and disadvantages of the two kinds of sea clutter suppression techniques are compared. The performance of sea clutter suppression technology based on AR model is better. The one-dimensional range profile of ship target is affected by sea clutter, which leads to the increase of inter-class convergence of features. The range profile recognition performance of ship target under low signal-to-clutter ratio can be improved.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52

【参考文献】