基于多重分形关联和深度CNN的常规雷达目标分类方法  

Target classification method of conventional radar based on multifractal correlation and deep CNN

作　　者：李秋生[1,2] 朱化娟 胡俊勇 LI Qiusheng;ZHU Huajuan;HU Junyong(Research Center of Intelligent Control Engineering Technology,Gannan Normal University,Ganzhou 341000,China;School of Physics and Electronic Information,Gannan Normal University,Ganzhou 341000,China)

机构地区：[1]赣南师范大学智能控制工程技术研究中心,江西赣州341000 [2]赣南师范大学物理与电子信息学院,江西赣州341000

出　　处：《安徽大学学报（自然科学版）》2023年第3期50-55,共6页Journal of Anhui University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61561004);江西省教育厅科学技术项目(GJJ201408);赣南师范大学研究生创新专项资金资助项目(YCX22A041)。

摘　　要：针对传统分类方法中飞机雷达回波信号识别分类精度低、人工定义特征稳定性差的问题,提出基于多重分形关联特征和深度卷积神经网络(convolutional neural network,简称CNN)的雷达目标分类方法.首先,对输入训练数据进行多重分形关联分析,将多重分形关联谱的投影图作为输入特征图;然后,利用深度卷积神经网络对特征进行训练,得到训练模型;最后,使用训练后的模型对目标进行分类.实验结果表明:相对于其他3种方法,该文方法有更强的飞机分类性能.Aiming at the problems of low accuracy of recognition and classification of aircraft radar echo signals in traditional classification methods and poor stability of manually defined features,a radar target classification method based on multifractal correlation features and convolutional neural network(CNN)was proposed.Firstly,the input training data was analyzed by multifractal correlation,and the projection of multifractal correlation spectrum was used as the input feature map.Then,the deep convolutional neural network was used to train the features,and the trained model was obtained.Finally,the trained model was used to identify the target.Experimental results showed that compared with the other three methods,the proposed method had stronger aircraft classification performance.

关 键 词：低分辨雷达 目标分类 多重分形关联谱 特征提取 深度卷积神经网络 

分 类 号：TN957.51[电子电信—信号与信息处理]