基于高次时频谱特征的LPI雷达信号识别  

LPI radar signal recognition based on high-order time-frequency spectrum features

作　　者：李世通 金小萍 孙杰[2] 汪晓锋 LI Shitong;JIN Xiaoping;SUN Jie;WANG Xiaofeng(Key Laboratory of Electromagnetic Wave Information Technology and Metrology of Zhejiang Province,College of Information Engineering,China Jiliang University,Hangzhou 310000,China;Zhejiang Institute of Metrology,Hangzhou 310000,China)

机构地区：[1]中国计量大学信息工程学院浙江省电磁波信息技术与计量检测重点实验室,杭州310000 [2]浙江省计量科学研究院,杭州310000

出　　处：《北京航空航天大学学报》2025年第1期314-320,共7页Journal of Beijing University of Aeronautics and Astronautics

基　　金：浙江省自然科学基金(LQ20F020021);浙江省电磁波信息技术与计量检测重点实验室开放式项目(2019KF0003)。

摘　　要：针对传统LPI雷达信号识别算法在低信噪比下识别率较低的问题,提出了一种基于高次时频特征的雷达信号识别算法。利用时频变换得到雷达信号的时频分布,时频谱做幂次化计算得到信号的高次时频图像,提取时频图像的灰度梯度共生矩阵和伪Zernike特征并组成联合特征向量,通过支持向量机实现雷达信号的分类识别。实验结果表明:在信噪比为−6 dB时,所提算法对8种典型雷达信号的整体识别准确率能达到95%以上。In view of the low recognition rate of traditional low probability of intercept(LPI)radar signal recognition algorithms under low signal-to-noise ratios,a radar signal recognition algorithm based on high-order timefrequency features was proposed.The proposed algorithm firstly obtained the time-frequency distribution of radar signals by time-frequency transform,and then the power calculation of the time-frequency spectrum was done to obtain the high-order time-frequency image of the signal.The gray gradient co-generation matrix and pseudo-Zernike features of the time-frequency image were extracted and formed into a joint feature vector,and finally,the classification recognition of the radar signal was realized by the support vector machine(SVM).The experimental results show that the overall recognition accuracy of the proposed algorithm can reach more than 95%for eight typical radar signals when the signal-to-noise ratio is−6 dB.

关 键 词：雷达信号识别 时频变换 高次时频 特征提取 支持向量机 

分 类 号：TN791[电子电信—电路与系统]