基于深度自编码多维特征融合的慢动目标检测

详细信息    查看全文 | 推荐本文 |

英文篇名：Slow Moving Target Detection Based on Multi-dimensional Feature Fusion Using Deep Autoencoder 作者：张文涛 ; 许治国 ; 郑霖 ; 杨超 英文作者：ZHANG Wentao;XU Zhiguo;ZHENG Lin;YANG Chao;School of Electronic Engineering and Automation,Guilin University of Electronic Technology;Guangxi Key Laboratory of Wireless Wideband Communication and Signal Processing,Guilin University of Electronic Technology; 关键词：目标检测 ; 深度自编码 ; 特征提取 ; 多维特征融合 ; 时频变换 ; 脉冲压缩 英文关键词：target detection;;deep autoencoder;;feature extraction;;multi-dimensional feature fusion;;time-frequency transform;;pulse compression 中文刊名：JSJC 英文刊名：Computer Engineering 机构：桂林电子科技大学电子工程与自动化学院;桂林电子科技大学广西无线宽带通信与信号处理重点实验室; 出版日期：2019-05-15 出版单位：计算机工程 年：2019 期：v.45;No.500 基金：国家自然科学基金(61371107,61565004,61571143);; 国家科技重大专项“高精度位置测量系统在线标定技术研究”(2017ZX02101007-003) 语种：中文; 页：JSJC201905023 页数：7 CN：05 ISSN：31-1289/TP 分类号：149-154+160

摘要

针对强杂波环境下慢动目标检测存在的多普勒频移低、杂波干扰强、特征提取困难等问题,提出一种多维特征融合的检测算法。利用时频变换和脉冲压缩解析回波信息,提取目标回波时频域和距离像的特征,将特征串联输入到深度自编码网络中进行融合。深度自编码网络通过自主学习提取目标不同维度的特征,增强多维特征联合检测性能。仿真结果表明,与直接利用单域特征的深度自编码以及利用SVM进行目标检测的算法相比,该算法能有效融合时频域与距离像特征,实现特征互补,提高目标检测的鲁棒性与识别精度。
        To address the problems of low Doppler frequency shift,strong clutter interference,difficult feature extraction in slow moving target detection under strong clutter environment,a target detection method based on multi-dimensional feature fusion is proposed.Firstly,the echo information is analyzed by using time-frequency transform and pulse compression,and the features of the target echo in the time-frequency domain and range image are extracted.Then,the features are connected in series into deep autoencoder network for fusion.Finally,the deep autoencoder network is used to extract different dimensions of target characteristics,and enhance the joint detection performance of multi-dimensional features based on autonomous learning.Simulation results show that compared with the target detection based on single domain feature using deep autoencoder and the target detection using SVM,the proposed method can fuse the features of the time-frequency domain and range image effectively,having advantages of both,and improve the robustness and recognition precision of target detection.

引文

[1]刘凯,林基明,郑霖等.基于深度自编码网络的慢速移动目标检测[J].计算机工程,2018,44(2):129-134.
    [2]陈小龙,关键,于晓涵等.基于短时稀疏时频分布的雷达目标微动特征提取及检测方法[J].电子与信息学报,2017,39(5):1017-1023.
    [3]LIANG Jing,LIANG Qilian.Sense-through-foliage target detec-tion using UWB radar sensor netw orks[J].Pattern Recognition Letters,2010,31(11):1412-1421.
    [4]JOKANOVIC B,AMIN M G,ZHANG Yimin,et al.M ulti-w indow time-frequency signature reconstruction from undersampled continuous-w ave radar measurements for fall detection[J].IET Radar Sonar and Navigation,2014,9(2):173-183.
    [5]ZHAI Shijun,JIANG Ting.A novel sense-through-foliage target recognition system based on sparse representation and improved particle swarm optimization-based support vector machine[J].Measurement,2013,46(10):3994-4004.
    [6]赵飞翔,刘永祥,霍凯.基于栈式降噪稀疏自动编码器的雷达目标识别方法[J].雷达学报,2017,6(2):149-156.
    [7]QI Yumei,SHEN Changqing,WANG Dong,et al.Stacked sparse autoencoder-based deep network for fault diagnosis of rotating machinery[J].IEEE Access,2017,5:15066-15079.
    [8]WU Qisong,ZHANG Yimin,TAO Wenbin,et al.Radarbased fall detection based on Doppler time-frequency signatures for assisted living[J].IET Radar Sonar and Navigation,2015,9(2):164-172.
    [9]JOKANOVIC B,AMIN M G,AHMAD F.Effect of data representations on deep learning in fall detection[C]//Proceedings of Sensor Array and M ultichannel Signal Processing Workshop.Washington D.C.,USA:IEEEPress,2016:1-5.
    [10]康妙,计科峰,冷祥光,等.基于栈式自编码器特征融合的SAR图像车辆目标识别[J].雷达学报,2017,6(2):167-176.
    [11]JOKANOVIC B,AMIN M.Fall detection using deep learning in range-Doppler radars[J].IEEE Transactions on Aerospace and Electronic Systems,2018,54(1):180-189.
    [12]HAYKIN S,BHATTACHARYA T K.Modular learning strategy for signal detection in a nonstationary environment[J].IEEETransactions on Signal Processing,1997,45(6):1619-1637.
    [13]EROL B,AMIN M G.Fall motion detection using combined range and Doppler features[C]//Proceedings of Signal Processing Conference.Washington D.C.,USA:IEEE Press,2016:2075-2080.
    [14]GROZDIC'D T,JOVICIC S T.Whispered speech recognition using deep denoising autoencoder and inverse filtering[J].IEEE/ACM Transactions on Audio Speech and Language Processing,2017,59(12):15-22.
    [15]牛玉虎,卷积稀疏自编码神经网络[J].计算机与现代化,2017(2):22-29.