基于卷积神经网络的海上微动目标检测与分类方法

苏宁远 陈小龙 关键 牟效乾 刘宁波

苏宁远, 陈小龙, 关键, 牟效乾, 刘宁波. 基于卷积神经网络的海上微动目标检测与分类方法[J]. 雷达学报, 2018, 7(5): 565-574. doi: 10.12000/JR18077
引用本文: 苏宁远, 陈小龙, 关键, 牟效乾, 刘宁波. 基于卷积神经网络的海上微动目标检测与分类方法[J]. 雷达学报, 2018, 7(5): 565-574. doi: 10.12000/JR18077
Su Ningyuan, Chen Xiaolong, Guan Jian, Mou Xiaoqian, Liu Ningbo. Detection and Classification of Maritime Target with Micro-motion Based on CNNs[J]. Journal of Radars, 2018, 7(5): 565-574. doi: 10.12000/JR18077
Citation: Su Ningyuan, Chen Xiaolong, Guan Jian, Mou Xiaoqian, Liu Ningbo. Detection and Classification of Maritime Target with Micro-motion Based on CNNs[J]. Journal of Radars, 2018, 7(5): 565-574. doi: 10.12000/JR18077

基于卷积神经网络的海上微动目标检测与分类方法

doi: 10.12000/JR18077
基金项目: 国家自然科学基金(61871391,61501487,61871392,U1633122,61471382,61531020);国防科技基金(2102024);山东省高校科研发展计划(J17KB139);泰山学者和中国科协青年人才托举工程(YESS20160115)专项经费
详细信息
    作者简介:

    苏宁远(1995–),男,山东烟台人,硕士在读。主要研究方向为智能雷达信号处理、目标检测。E-mail: 965291799@qq.com

    陈小龙(1985–),男,山东烟台人,博士,讲师。研究领域包括雷达动目标检测、海杂波抑制、雷达信号精细化处理等。入选中国科协“青年人才托举工程”,获中国电子学会优秀博士学位论文奖,第十九届中国专利优秀奖,中国电子学会科技进步三等奖。E-mail: cxlcxl1209@163.com

    关 键(1968–),男,辽宁锦州人,教授,博士生导师。主要研究方向为包括雷达目标检测与跟踪、侦察图像处理和信息融合。获国家科技进步二等奖1项、军队科技进步一等奖2项,山东省技术发明一等奖1项;“百千万人才工程”国家级人选,入选教育部新世纪优秀人才支持计划。E-mail: guanjian_68@163.com

    牟效乾(1995–),男,山东烟台人,硕士在读。研究领域包括智能雷达信号处理、动目标检测等。E-mail: 1012226010@qq.com

    刘宁波(1983–),男,山东烟台人,博士,讲师,研究方向为雷达信号处理、海杂波抑制与目标智能检测。E-mail: lnb198300@163.com

    通讯作者:

    陈小龙  cxlcxl1209@163.com

Detection and Classification of Maritime Target with Micro-motion Based on CNNs

Funds: The National Natural Science Foundation of China (61871391, 61501487, 61871392, U1633122, 61471382, 61531020), National Defense Science Foundation (2102024), Scientific Research Development of Shandong (J17KB139), Special Funds of Taishan Scholars of Shandong and Young Elite Scientist Sponsorship Program of CAST (YESS20160115)
  • 摘要: 该文利用深度学习的高维特征泛化学习能力,将卷积神经网络(CNN)用于海上目标微多普勒的检测和分类。首先,在海面微动目标模型的基础上,在实测海杂波背景中分别构建4种类型微动信号的2维时频图,并作为训练和测试数据集;然后,分别采用LeNet, AlexNet和GoogLeNet 3种CNN模型进行二元检测和多种微动类型分类,并进行比较,研究信杂比对检测和分类性能的影响。最后,与传统的支持向量机方法进行比较,结果表明,所提方法能够智能学习微动特征,具有更好的检测和分类性能,可为杂波背景下的雷达动目标检测和识别提供新的技术途径。

     

  • 图  1  本文采用的3种CNN网络结构

    Figure  1.  Structure of three CNNs used in this paper

    图  2  CNN中各卷积层的数据特征(以LeNet为例)

    Figure  2.  Data characteristics of each convolution layer in CNN (e.g., LeNet)

    图  3  所提方法处理流程图

    Figure  3.  Processing flow diagram of the proposed method

    图  4  数据集示例

    Figure  4.  Examples of datasets

    图  5  错判示例图

    Figure  5.  Examples of wrong judgment

    图  7  不同模型目标分类结果

    Figure  7.  Results of target classification via different models

    图  6  不同信杂比下目标检测分类性能

    Figure  6.  Performances of target classification under different SCRs

    表  1  微动参数设置

    Table  1.   Configuration of micro-motion parameters

    运动类型 初速度
    (m/s)
    加速度
    (m/s2)
    急动度
    (m/s3)
    采样点数 运动类型 角速度 微动周期(s) 采样点数
    匀加速 [5, 15] [–16, 16] 211 Hz, 0.50 s 微动Ⅰ [16] $\left| {{{\bar \omega }_{x{\rm m}}}} \right|$=[0.34, 0.42] rad/s
    $\left| {{{\bar \omega }_{y{\rm m}}}} \right|$=[0.15, 0.17] rad/s
    $\left| {{{\bar \omega }_{z{\rm m}}}} \right|$=[0.07, 0.09] rad/s
    ${T_x}$=26.4
    ${T_y}$=11.2
    ${T_z}$=33.0
    28 Hz, 8 s
    非匀变速 [50, 300] [–160, 160] [–160, 160] 214 Hz, 0.25 s 微动Ⅱ [16] $\left| {{{\bar \omega }_{x{\rm m}}}} \right|$=[0.61, 0.65] rad/s
    $\left| {{{\bar \omega }_{y{\rm m}}}} \right|$=[0.95, 1.07] rad/s
    $\left| {{{\bar \omega }_{z{\rm m}}}} \right|$=[0.52, 0.56] rad/s
    ${T_x}$=12.2
    ${T_y}$=6.7
    ${T_z}$=14.2
    210 Hz, 8 s
    下载: 导出CSV

    表  2  CNN模型训练时长

    Table  2.   Training time of CNN models

    模型 目标检测模型 目标分类模型
    LeNet AlexNet GoogLeNet LeNet AlexNet GoogLeNet
    训练用时(min) 68.00 37.52 175.00 27.92 38.40 54.43
    下载: 导出CSV

    表  3  不同模型目标检测结果

    Table  3.   Results of target detection of different models

    模型 LeNet AlexNet GoogLeNet
    虚警概率 1.24% 0.04% 0.24%
    检测概率 92.28% 84.44% 90.94%
    仿真时间(min) 41.92 57.15 43.77
    下载: 导出CSV

    表  4  基于CNN和SVM的海面目标检测与分类结果比较

    Table  4.   Comparison of maritime targets detection and classification results based on CNN and SVM

    分类方法 虚警概率 检测概率(%) 识别概率(SCR=–4 dB)(%)
    SCR=4 dB SCR=–4 dB SCR=–12 dB SCR=–20 dB 匀变速 变加速 微动Ⅰ 微动Ⅱ
    CNN 0.0290 95.80 96.30 94.11 90.62 100.00 100.00 99.65 100.00
    SVM 0.2207 80.52 79.90 77.81 77.31 84.30 84.46 73.56 94.25
    下载: 导出CSV
  • [1] Darzikolaei M A, Ebrahimzade A, and Gholami E. Classification of radar clutters with artificial neural network[C]. Proceedings of the 2nd International Conference on Knowledge-Based Engineering and Innovation, Tehran, Iran, 2015: 577–581.
    [2] 陈小龙, 关键, 何友. 微多普勒理论在海面目标检测中的应用及展望[J]. 雷达学报, 2013, 2(1): 123–134. DOI: 10.3724/SP.J.1300.2013.20102

    Chen Xiao-long, Guan Jian, and He You. Applications and prospect of micro-motion theory in the detection of sea surface target[J].Journal of Radars, 2013, 2(1): 123–134. DOI: 10.3724/SP.J.1300.2013.20102
    [3] 罗迎, 张群, 王国正, 等. 基于复图像OMP分解的宽带雷达微动特征提取方法[J]. 雷达学报, 2012, 1(4): 361–369. DOI: 10.3724/SP.J.1300.2012.20065

    Luo Ying, Zhang Qun, Wang Guo-zheng, et al. Micro-motion signature extraction method for wideband radar based on complex image OMP decomposition[J]. Journal of Radars, 2012, 1(4): 361–369. DOI: 10.3724/SP.J.1300.2012.20065
    [4] Chen X L, Guan J, Bao Z H, et al. Detection and extraction of target with micromotion in spiky sea clutter via short-time fractional Fourier transform[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014, 52(2): 1002–1018. DOI: 10.1109/TGRS.2013.2246574
    [5] Chen X L, Guan J, Li X Y, et al. Effective coherent integration method for marine target with micromotion via phase differentiation and radon-Lv’s distribution[J]. IET Radar,Sonar&Navigation, 2015, 9(9): 1284–1295. DOI: 10.1049/iet-rsn.2015.0100
    [6] Wagner S A. SAR ATR by a combination of convolutional neural network and support vector machines[J]. IEEE Transactions on Aerospace and Electronic Systems, 2016, 52(6): 2861–2872. DOI: 10.1109/TAES.2016.160061
    [7] 田壮壮, 占荣辉, 胡杰民, 等. 基于卷积神经网络的SAR图像目标识别研究[J]. 雷达学报, 2016, 5(3): 320–325. DOI: 10.12000/JR16037

    Tian Zhuang-zhuang, Zhan Rong-hui, Hu Jie-min, et al. SAR ATR based on convolutional neural network[J]. Journal of Radars, 2016, 5(3): 320–325. DOI: 10.12000/JR16037
    [8] Kim Y and Toomajian B. Hand gesture recognition using micro-Doppler signatures with convolutional neural network[J]. IEEE Access, 2016, 4: 7125–7130. DOI: 10.1109/ACCESS.2016.2617282
    [9] 王俊, 郑彤, 雷鹏, 等. 深度学习在雷达中的研究综述[J]. 雷达学报, 2018, 7(4): 395–411. DOI: 10.12000/JR18040

    Wang Jun, Zheng Tong, Lei Peng, et al. Survey of study on deep learning in radar[J]. Journal of Radars, 2018, 7(4): 395–411. DOI: 10.12000/JR18040
    [10] 徐彬, 陈渤, 刘宏伟, 等. 基于注意循环神经网络模型的雷达高分辨率距离像目标识别[J]. 电子与信息学报, 2016, 38(12): 2988–2995. DOI: 10.11999/JEIT161034

    Xu Bin, Chen Bo, Liu Hong-wei, et al. Attention-based recurrent neural network model for radar high-resolution range profile target recognition[J]. Journal of Electronics&Information Technology, 2016, 38(12): 2988–2995. DOI: 10.11999/JEIT161034
    [11] 王星, 周一鹏, 周冬青, 等. 基于深度置信网络和双谱对角切片的低截获概率雷达信号识别[J]. 电子与信息学报, 2016, 38(11): 2972–2976. DOI: 10.11999/JEIT160031

    Wang Xing, Zhou Yi-peng, Zhou Dong-qing, et al. Research on low probability of intercept radar signal recognition using deep belief network and bispectra diagonal slice[J]. Journal of Electronics&Information Technology, 2016, 38(11): 2972–2976. DOI: 10.11999/JEIT160031
    [12] 徐真, 王宇, 李宁, 等. 一种基于CNN的SAR图像变化检测方法[J]. 雷达学报, 2017, 6(5): 483–491. DOI: 10.12000/JR17075

    Xu Zhen, Wang Robert, Li Ning, et al. A novel approach to change detection in SAR images with CNN classification[J]. Journal of Radars, 2017, 6(5): 483–491. DOI: 10.12000/JR17075
    [13] 徐丰, 王海鹏, 金亚秋. 深度学习在SAR目标识别与地物分类中的应用[J]. 雷达学报, 2017, 6(2): 136–148. DOI: 10.12000/JR16130

    Xu Feng, Wang Hai-peng, and Jin Ya-qiu. Deep learning as applied in SAR target recognition and terrain classification[J]. Journal of Radars, 2017, 6(2): 136–148. DOI: 10.12000/JR16130
    [14] 陈小龙, 董云龙, 李秀友, 等. 海面刚体目标微动特征建模及特性分析[J]. 雷达学报, 2015, 4(6): 630–638. DOI: 10.12000/JR15079

    Chen Xiao-long, Dong Yun-long, Li Xiu-you, et al. Modeling of micromotion and analysis of properties of rigid marine targets[J]. Journal of Radars, 2015, 4(6): 630–638. DOI: 10.12000/JR15079
    [15] Prusa J D and Khoshgoftaar T M. Improving deep neural network design with new text data representations[J]. Journal of Big Data, 2017, 4: 7. DOI: 10.1186/s40537-017-0065-8
    [16] 高建军. 多径和海杂波干扰下的舰船ISAR成像及横向定标[D]. [博士论文], 哈尔滨工业大学, 2010.

    Gao Jian-jun. ISAR ship imaging and cross-range scaling with multipath and sea clutter interference[D]. [Ph.D. dissertation], Harbin Institute of Technology, 2010.
  • 加载中
图(7) / 表(4)
计量
  • 文章访问数:  5768
  • HTML全文浏览量:  1522
  • PDF下载量:  932
  • 被引次数: 0
出版历程
  • 收稿日期:  2018-09-14
  • 修回日期:  2018-10-16
  • 网络出版日期:  2018-10-28

目录

    AI智能化辅读

    AI智能问答

        /

        返回文章
        返回