基于时频分布图像和主分量分析的脉内调制识别算法研究  

Study on intra-pulse modulation recognition method based on time-frequency distribution image and principal component analysis

作　　者：冀贞海[1,2] 朱伟强[1] 赵力[2] 

机构地区：[1]中国航天科工集团8511研究所,江苏南京210007 [2]东南大学信息科学与工程学院,江苏南京210096

出　　处：《电路与系统学报》2009年第2期22-26,共5页Journal of Circuits and Systems

基　　金：国家自然科学基金(60472058);教育部博士点基金(20050286001);教育部新世纪优秀人才支持计划(NCET-04-0483)

摘　　要：本文提出了一种利用模式识别中图像处理的方法来进行雷达脉内调制特征识别的新算法。通过对雷达信号时频分布图像进行二维小波分解,并对其进行主分量分析,获得每类雷达信号中的特征参数,构建相应的分类器,并利用此分类器对信号进行识别。仿真结果表明,利用信号的时频分布图像来进行信号识别是完全可行的,提取的特征参数具有很好的鲁棒性,具有较强的抗噪声能力,同时可以取得较高识别率。This paper proposes a new method for radar intra-pulse modulation characteristic recognition by means of image processing method in the field of pattern recognition. Through two-dimensional wavelet decomposition for time-frequency distribution image and principal component analysis, the feature parameters of each type radar signal can be obtained, and constructed the corresponding classifiers, which can be used for signals identification. The experiment results show it is feasible completely for signals recognition using time-frequency image of signals, and the extracted feature parameters have good performance, and also have higher recognition rate.

关 键 词：时频分布图像 主分量分析 脉内调制识别 

分 类 号：TN974[电子电信—信号与信息处理]