基于EMD-MDGAN的HRRP增扩方法

doi:10.19678/j.issn.1000-3428.0061223

摘要/Abstract

摘要： 高分辨率距离像（HRRP）是弹道目标识别的主要特征，由于其为非合作目标因此观测频率极低，导致带标签样本量严重不足，而混合密度生成对抗网络（MDGAN）作为生成HRRP的有效方法，存在模式崩溃、网络不易收敛等问题。提出一种基于误差匹配分布（EMD）改进MDGAN的弹道目标HRRP增扩方法EMD-MDGAN。将生成器、残差网络和注意力机制相结合，通过残差结构解决梯度消失的问题，利用注意力机制提高生成器中自编码器的特征提取能力，并把误差匹配思想引入损失函数设计中，以增强模型的稳定性，使网络更易收敛。实验结果表明，该模型在有效解决模式崩溃问题的基础上，可缩小生成样本与真实样本分布间差异，生成具有一定真实性、可靠性、多样性的数据，实现HRRP数据增扩。

关键词: 混合密度生成对抗网络, 残差网络, 注意力机制, 误差匹配分布, 高分辨率距离像

Abstract: High-Resolution Response Profile(HRRP) is one of the main features for ballistic target recognition, but it is a non-cooperative target and thus limited by the low observation frequency, leading to a lack of labeled samples. Mixture Density Generative Adversarial Network(MDGAN) is an effective way to generate HRRP, but deeply suffers from mode collapse and difficulties in network convergence.To address the problems, a new HRRP expansion method named EMD-MDGAN is proposed for trajectory targets.EMD-MDGAN is based on Mixture Density Generative Adversarial Network(MDGAN), which is improved by Error Matching Distribution(EMD).The method employs the residual network structure to solve the problem of gradient disappearance, and uses the attention mechanism to improve the feature extraction ability of the auto-encoder in the generator.On this basis, the idea of error matching is introduced into the design of loss function to enhance the stability of the model and make the network more convergent.The experimental results show that the model can avoid mode collapse, and thus reduce the distribution difference between the generated samples and actual samples.It can ensure the authenticity, reliability and diversity of generated data, and realize HRRP data expansion.

Key words: Mixture Density Generative Adversarial Network(MDGAN), residual network, attention mechanism, error matching distribution, High-Resolusion Response Profile(HRRP)

中图分类号:

TP391

王紫娇, 王晓丹. 基于EMD-MDGAN的HRRP增扩方法[J]. 计算机工程, 2021, 47(9): 259-265.

WANG Zijiao, WANG Xiaodan. HRRP Expansion Method Based on EMD-MDGAN[J]. Computer Engineering, 2021, 47(9): 259-265.

http://www.ecice06.com/CN/Y2021/V47/I9/259

图/表 9

20210917192613

20210917192617

20210917192624

20210917192628

20210917192632

20210917192636

20210917192639

20210917192643

20210917192647

参考文献

[1] 余宪.基于生成对抗网络的雷达高分辨距离像生成方法研究[D].厦门:厦门大学, 2019. YU X.A study on radar High-Resolution Response Profile(HRRP) generation method based on generating confrontation network[D].Xiamen:Xiamen University, 2019.(in Chinese)
[2] WANG J J, LIU Z, XIE R, et al.Radar HRRP target recognition based on dynamic learning with limited training data[J].Remote Sensing, 2021, 13(4):750.
[3] 郭泽坤, 田隆, 韩宁, 等.采用CNN-SSD的雷达HRRP小样本目标识别方法[J].西安电子科技大学学报, 2021, 48(2):7-14. GUO Z K, TIAN L, HAN N, et al.Radar HRRP based few-shot target recognition with CNN-SSD[J].Journal of Xi'an University of Electronic Science and Technology, 2021, 48(2):7-14.(in Chinese).
[4] GOODFELLOW I J, POUGET-ABADIE J, MIRZA M, et al.Generative adversarial networks[EB/OL].[2021-02-10].https://arxiv.org/pdf/1406.2661.pdf.
[5] 苏炯铭, 刘鸿福, 项凤涛, 等.深度神经网络解释方法综述[J].计算机工程, 2020, 46(9):1-15. SU J M, LIU H F, XIANG F T, et al.Survey of interpretation methods for deep neural networks[J]. Computer Engineering, 2020, 46(9):1-15.(in Chinese)
[6] 柴梦婷, 朱远平.生成式对抗网络研究与应用进展[J]. 计算机工程, 2019, 45(9):222-234. CHAI M T, ZHU Y P.Research and application progress of generative adversarial networks[J].Computer Engineering, 2019, 45(9):222-234.(in Chinese)
[7] 宋益恒, 王彦华, 李阳, 等.基于深度生成网络的雷达HRRP生成技术[J].信号处理, 2019, 35(6):1118-1122. SONG Y H, WANG Y H, LI Y, et al.Radar HRRP generation technology based on depth generation network[J].Signal Processing, 2019, 35(6):1118-1122.(in Chinese)
[8] MAO X, LI Q, XIE H, et al.Least squares generative adversarial networks[C]//Proceedings of 2017 IEEE International Conference on Computer Vision.Washington D.C., USA:IEEE Press, 2016:2813-2821.
[9] KARRAS T, AILA T, LAINE S, et al.Progressive growing of GANs for improved quality, stability, and variation[EB/OL].[2021-02-10].https://arxiv.org/pdf/1710. 10196.pdf.
[10] HUANG X, LI Y, POURSAEED O, et al.Stacked generative adversarial networks[EB/OL].[2021-02-10].https://arxiv.org/pdf/1612.04357v4.pdf.
[11] SHAHAM T R, DEKEL T, MICHAELI T.SinGAN:learning a generative model from a single natural image[C]//Proceedings of 2019 IEEE International Conference on Computer Vision.Washington D.C., USA:IEEE Press:2019:4569-4579.
[12] EGHBAL-ZADEH H.Mixture density generative adversarial networks[EB/OL].[2021-02-10].https://arxiv.org/pdf/1811.00152.pdf.
[13] MONTUFAR G F, PASCANU R, CHO K, et al.On the number of linear regions of deep neural networks[C]//Proceedings of NIPS'14.Vancouver, Canada:[s.n.], 2014:2924-2932.
[14] HE K M, ZHANG X, REN S, et al.Deep residual learning for image recognition[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Ppattern Recognition.Washington D.C., USA:IEEE Press, 2016:770-778.
[15] 颜贝.基于对抗式网络的图像数据生成技术研究[D].成都:中国科学院光电技术研究所, 2020. YAN B.A study on image data generation technology based on confrontational network[D].Chengdu:Institute of Optoelectronic Technology, Chinese Academy of Sciences, 2020.(in Chinese)
[16] BERTHELOT D, SCHUMM T, METZ L.BEGAN:boundary equilibrium generative adversarial networks[EB/OL].(2017-05-31)[2021-02-10].https://arxiv.org/pdf/1703.10717.pdf.
[17] 向前, 王晓丹, 李睿, 等.基于DCNN的弹道中段目标HRRP图像识别[J].系统工程与电子技术, 2020, 42(11):2426-2433. XIANG Q, WANG X D, LI R, et al.HRRP image recognition based on DCNN ballistic midcourse target[J].Systems Engineering and Electronic Technology, 2020, 42(11):2426-2433.(in Chinese)
[18] BARRATT S, SHARMA R.A note on the inception score[EB/OL].[2021-02-10].https://arxiv.org/pdf/1801.01973.pdf.
[19] HEUSEL M, RAMSAUER H, UNTERTHINER T, et al.GANs trained by a two time-scaleupdate rule converge to a local Nash equilibrium[C]//Proceedings of NIPS'17.Long Beach, USA:[s.n.], 2017:6626-6637.
[20] BINKOWSKI M, SUTHERLAND D J, ARBEL M, et al.Demystifying MMD GANs[EB/OL].[2021-02-10].https://www.researchgate.net/publication/322306034_Demystifying_MMD_GANs.
[21] RADFORD A, METZ L, CHINTALA S.Unsupervised representation learning with deep convolutional generative adversarial networks[EB/OL].[2021-02-10].https://arxiv.org/pdf/1511.06434.pdf.
[22] ARJOVSKY M, CHINTALA S, BOTTOU L E O.Wasserstein GAN[EB/OL].[2021-02-10].https://arxiv.org/pdf/1701.07875.pdf.
[23] GULRAJANI I, AHMED F, ARJOVSKY M, et al.Improved training of Wasserstein GANs[EB/OL].[2021-02-10].https://arxiv.org/pdf/1704.00028.pdf.
[24] MAO X, LI Q, XIE H, et al.Least squares generative adversarial networks[EB/OL].[2021-02-10].https://arxiv.org/pdf/1611.04076.pdf.

选择文件类型/文献管理软件名称

选择包含的内容