基于雷达高分辨距离像序列识别的研究综述  

A Survey of Recognition Research Based on Radar HRRP Sequence

作　　者：卢中原 王中训[1] 但波 Lu Zhongyuan;Wang Zhongxun;Dan Bo(School of Physics and Electronic Information,Yantai University,Yantai 264005,China;College of Coast Guard,Naval Aviation University,Yantai 264001,China)

机构地区：[1]烟台大学物理与电子信息学院,山东烟台264005 [2]海军航空大学岸防兵学院,山东烟台264001

出　　处：《兵工自动化》2022年第9期65-71,共7页Ordnance Industry Automation

摘　　要：为了解雷达目标更多的详细信息,对基于高分辨距离像(high resolution range profile,HRRP)序列的目标识别方法进行分析。对隐马尔可夫模型(hidden Markov model,HMM)、卷积神经网络(convolutional neural network,CNN)、循环神经网络(recurrent neural network,RNN)和长短期记忆网络(long short-term memory,LSTM)等面向序列识别的分类器进行阐述和讨论,分析不同分类器用于HRRP目标识别的发展历程,指出不同识别方法的优缺点及适用性条件。结果表明,该研究可为不同识别场景下应用合适的分类器提供一些思路。In order to get more detailed information of radar target,the target recognition method based on high resolution range profile(HRRP)sequence is analyzed.A hidden Markov model(HMM),a convolution neural network(CNN),a recurrent neural network(RNN),and long short-term memory(LSTM),analyzes the development process of different classifiers for HRRP target recognition,and points out the advantages,disadvantages and applicability conditions of different recognition methods.The results show that this study can provide some ideas for the application of appropriate classifiers in different recognition scenarios.

关 键 词：高分辨距离像 雷达 序列识别 人工智能 雷达自动目标识别 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]