基于联合时频特征和HMM的多方位SAR目标识别

Journal of Systems Engineering and Electronics ›› 2010, Vol. 32 ›› Issue (4): 712-717.

基于联合时频特征和HMM的多方位SAR目标识别

张新征, 黄培康

(目标与环境电磁散射辐射国防科技重点实验室，北京 100854)

出版日期:2010-04-23 发布日期:2010-01-03

Multi-aspect SAR target recognition based on combined time-frequency feature and HMM

ZHANG Xin-zheng, HUANG Pei-kang

(National Electromagnetic Scattering Laboratory, Beijing 100854, China)

Online:2010-04-23 Published:2010-01-03

摘要/Abstract

摘要：

研究了联合时频特征和隐马尔科夫模型(hidden Markov model, HMM)的多方位合成孔径雷达(synthetic aperture radar, SAR)目标识别方法。利用HMM模型可以有效地对多方位SAR目标特征分析及识别。在HMM多方位SAR目标识别中的关键之一是SAR目标回波高分辨率距离像(high resolution range profile, HRRP)的特征提取。提出了一种时变频因子加权Fisher鉴别的特征提取方法。利用MSTAR实测SAR目标数据集进行了特征提取和识别实验，实验结果验证了方法的有效性。

Abstract:

Multi-aspect sythetic aperture radar (SAR) target recognition based on combined time-frequency feature and hidden Markov model (HMM) is investigated. HMM is a powerful tool to analyze and recognize the characteristics of multi-aspect SAR targets as a framework. One of the critical technique is feature extraction from the high resolution range profile (HRRP) of target echoes in the framework. A time-varying frequency factor weighted Fisher discrimination time-frequency spectra feature extraction method is proposed. Recognition experiments are performed by the feature extraction method and HMM, which shows that the performance of this feature extraction method is effective.

张新征, 黄培康. 基于联合时频特征和HMM的多方位SAR目标识别[J]. Journal of Systems Engineering and Electronics, 2010, 32(4): 712-717.

ZHANG Xin-zheng, HUANG Pei-kang. Multi-aspect SAR target recognition based on combined time-frequency feature and HMM[J]. Journal of Systems Engineering and Electronics, 2010, 32(4): 712-717.