基于模糊-粗糙集和傅里叶-梅林变换的HRRP识别  

HRRP Recognition Based on Fuzzy-rough Set and Fourier-Mellin Transform

作　　者：付强[1] 胡磊[1] 周剑雄[1] 

机构地区：[1]国防科技大学ATR重点实验室,长沙410073

出　　处：《现代雷达》2009年第4期61-64,共4页Modern Radar

基　　金：国家安全自大基础研究(973)项目

摘　　要：提出一种基于模糊-粗糙集理论和傅里叶-梅林变换的高分辨雷达距离像识别方法。模糊-粗糙集理论将模糊集概念引入到粗糙集理论中,克服了传统粗糙集理论不能直接处理连续值特征的缺陷;傅里叶-梅林变换具有时移与尺度伸缩不变性,能够克服距离单元的平移、伸缩给识别带来的困难。实验表明,该方法具有强大的特征压缩能力,所设计识别系统具有良好的识别率。A new approach employing fuzzy-rough set theory and Fourier-Mellin transformation to solve HRRP recognition problem is presented in this paper. Fuzzy-rough set theory integrates fuzzy set concept with rough set theory so that it overcomes the incom- petence of rough set theory in dealing with continuous value features. Fourier-Mellin transform has the property of shift and scaling invariance, thus it can overcome the recognition difficulty caused by range bins shifting and scaling. It is demonstrated by simula- tion that the method is very powerful for feature compression and the designed recognition system works well with high probability of right classification.

关 键 词：距离像 雷达目标识别 模糊-粗糙集 傅里叶-梅林变换 特征压缩 C4.5分类器 

分 类 号：TN957.52[电子电信—信号与信息处理]