海杂波频谱的多重分形特性分析  

Multifractal property of sea clutter frequency spectrum

作　　者：刘宁波[1] 关键[1] 宋杰[1] 黄勇[1] 何友[1] 

机构地区：[1]海军航空工程学院电子信息工程系,烟台264001

出　　处：《中国科学：信息科学》2013年第6期768-783,共16页Scientia Sinica(Informationis)

基　　金：国家自然科学基金(批准号:61179017;60802088);"泰山学者"建设工程专项经费;航空科学基金(批准号:20095184004)资助项目

摘　　要：本文主要研究海杂波频谱的多重分形特性并分析其影响因素.首先介绍了将多重分形理论引入到海杂波频谱分析中的理论基础,然后,采用X波段与S波段相参雷达实测数据验证海杂波频谱的多重分形特性,并采用多重分形去趋势起伏分析(MF-DFA)法对海杂波频谱进行多重分形特性分析,分析结果表明海杂波频谱是一种受长程相关性和概率分布共同影响的多重分形序列,并且频域多重分形参数—广义Hurst指数对海杂波单元与目标单元具有一定的区分能力.此外,通过分析广义Hurst指数的影响因素还发现FFT时所采用的时间序列长度对广义Hurst指数影响相对较大,而FFT点数对其影响相对较小.This paper focuses on the multifractal property of the sea clutter frequency spectrum and analyzes the influencing factors of the multifractal property. First, the theoretical foundation introducing multifractal theory for analyzing the sea clutter frequency spectrum is laid. The datasets from the X-band and S-band coherent radars are then used to verify the multifractal property of the sea clutter frequency spectrum, and multifractal detrended fluctuation analysis （MF-DFA） is subsequently adopted for carrying out the multifractal analysis of the sea clutter frequency spectrum. The analytical results of the multifractal analysis show that the sea clutter frequency spectrum sequence is a multifractal sequence because of the long-range correlations and broad probability distributions. The multifractal parameter in the frequency domain, i.e. the generalized Hurst exponent in the frequency domain, has, to some extent, the ability to distinguish sea clutter range bins from target range bins. In addition, by analyzing the influencing factors of the generalized Hurst exponent in the frequency domain, it is found that the length of the time sequence used for the FFT has more affect on the generalized Hurst exponent in the frequency domain, while the length of the FFT has less affect.

关 键 词：海杂波 频谱 多重分形 自相似 去趋势起伏分析 广义Hurst 指数 

分 类 号：TN957.51[电子电信—信号与信息处理]