基于生成对抗神经网络的雷达遥感数据增广方法

doi:10.16182/j.issn1004731x.joss.20-0953

摘要/Abstract

摘要：

在雷达探测领域，数据样本无论在完备性还是多样性上，均不能满足深度学习模型有效训练的要求，模型极易出现过拟合现象，从而限制了相关技术在雷达探测领域的广泛应用。面向雷达探测领域的智能化应用需求，针对雷达数据样本不足问题，提出基于生成对抗神经网络的微波成像体制雷达数据增广方法。针对雷达数据样本特征不显著问题，结合标签平滑正则化方法，实现增广数据样本的自动标注，通过构建增广样本与真实样本协同的深度学习模型训练框架，实现模型在小规模雷达数据样本集上的鲁棒训练。基于公开雷达探测数据集，验证了该方法的有效性。

关键词: 深度学习, 雷达遥感探测, 生成对抗神经网络, 数据增广

Abstract:

In the research field of radar remote sensing, both the completeness and diversity of radar data samples cannot meet the requirement of effective training of deep learning models, and the models are prone to over-fitting, which significantly limits the wide application of deep learning techniques in this field. Targeting on the needs of intelligent application in radar remote sensing, a microwave imaging radar suited data augmentation method is proposed to solve the issue of insufficient radar data samples by leveraging the general framework of generative adversarial network. Aiming at the features of radar samples being not obvious, the label smoothing regularization technique is utilized to automatically classify the augmentated radar samples. The augmentated samples together with the real samples are collaboratively used to implement the robust training of deep learning models. The proposed method is verified by the experiments based on the extensive open-sourse radar remote sensing data.

Key words: deep learning, radar remote sensing, generative adversarial network, data augmentation

中图分类号:

TP391.9

康旭, 张晓峰. 基于生成对抗神经网络的雷达遥感数据增广方法[J]. 系统仿真学报, 2022, 34(4): 920-927.

Xu Kang, Xiaofeng Zhang. Radar Remote Sensing Data Augmentation Method Based on Generative Adversarial Network[J]. Journal of System Simulation, 2022, 34(4): 920-927.

图/表 9

图1

图2

图3

表1

图4

表2

表3

表4

图5

参考文献 20

1	赵健, 张智, 李帅. 雷达电磁环境仿真研究[J]. 舰船电子对抗, 2009, 32(5): 71-74.
	Zhao Jian, Zhang Zhi, Li Shuai. Study of Electromagnetic Environment Simulation for Radar[J]. Shipboard Electronic Countermeasure, 2009, 32(5): 71-74.
2	张文辉, 张永宁, 田雨书, 等. 基于SystemVue的复杂电磁环境中雷达半实物仿真[J]. 火控雷达技术, 2018, 47(2): 85-90.
	Zhang Wenhui, Zhang Yongning, Tian Yushu, et al. System Vue Based Hardware-in-loop Simulation of Radar in Complicated Electromagnetic Environment[J]. Fire Control Radar Technology, 2018, 47(2): 85-90.
3	Goodfellow I, Pouget-Abadie J, Mirza M, et al. Generative Adversarial Nets[C]// Advances in Neural Information Processing Systems. Montreal: MIT Press, 2014: 2672-2680.
4	Radford A, Metz L, Chintala S, et al. Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks[C]// International Conference on Learning Representations. San Juan, Puerto Rico: ICLR, 2016: 1-16.
5	Adler J, Lunz S. Banach Wasserstein Gan[C]// Advances in Neural Information Processing Systems. Montréal: MIT Press, 2018: 6754-6763.
6	Zhu J Y, Park T, Isola P, et al. Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks [C]// IEEE International Conference on Computer Vision. Venice: IEEE Computer Society, 2017: 2223-2232.
7	Hussein S A, Tirer T, Giryes R. Image-Adaptive GAN Based Reconstruction[C]// AAAI Conference on Artificial Intelligence. New York: AAAI, 2020: 3121-3129.
8	Kumar K, Kumar R, de Boissiere T, et al. Melgan: Generative Adversarial Networks for Conditional Waveform Synthesis[C]// Advances in Neural Information Processing Systems. Vancouver: MIT Press, 2019: 14910-14921.
9	Yang S, Xie L, Chen X, et al. Statistical Parametric Speech Synthesis Using Generative Adversarial Networks Under a Multi-task Learning Framework[C]// 2017 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU). Okinawa: IEEE, 2017: 685-691.
10	Zhang Y, Gan Z, Fan K, et al. Adversarial Feature Matching for Text Generation[C]// 34th International Conference on Machine Learning Learning. Sydney: ACM, 2017: 4006-4015.
11	Yu L, Zhang W, Wang J, et al. Seqgan: Sequence Generative Adversarial Nets with Policy Gradient[C]// Thirty-first AAAI Conference on Artificial Intelligence. San Francisco: AAAI, 2017.
12	Zheng Z, Zheng L, Yang Y, et al. Unlabeled Samples Generated by GAN Improve the Person Re-Identification Baseline in Vitro[C]// International Conference on Computer Vision. Venice: IEEE, 2017: 3774-3782.
13	Hou J, Zeng H, Cai L, et al. Multi-Label Learning with Multi-Label Smoothing Regularization for Vehicle Re-Identification[J]. Neurocomputing (S0925-2312), 2019, 345: 15-22.
14	Zheng Z, Zheng L, Yang Y. Unlabeled Samples Generated by Gan Improve the Person Re-identification Baseline in Vitro[C]// IEEE International Conference on Computer Vision. Venice: IEEE, 2017: 3754-3762.
15	Sun Y, Liu Z, Todorovic S, et al. Adaptive Boosting for SAR Automatic Target Recognition[J]. IEEE Transactions on Aerospace and Electronic Systems(S0018-9251), 2007, 43(1): 112-125.
16	Zhao Q, Principe J C. Support Vector Machines for SAR Automatic Target Recognition[J]. IEEE Transactions on Aerospace and Electronic Systems (S0018-9251), 2001, 37(2): 643-654.
17	Lecun Y, Bottou L. Gradient-Based Learning Applied to Document Recognition[J]. IEEE (S0018-9219), 1998, 86(11): 2278-2324.
18	Chen S, Wang H, Xu F, et al. Target Classification Using the Deep Convolutional Networks for SAR Images[J]. IEEE Transactions on Geoscience and Remote Sensing(S0196-2892), 2016, 54(8): 4806-4817.
19	Gulrajani I, Ahmed F, Arjovsky M, et al. Improved Training of Wasserstein Gans[C]// Advances in Neural Information Processing Systems. Long Beach: MIT Press, 5767-5777.
20	Mao X, Li Q, Xie H, et al. Least Squares Generative Adversarial Networks[C]// IEEE International Conference on Computer Vision. Venice: IEEE, 2017: 2794-2802.

目标类别	俯仰角15°	俯仰角17°
合计	2 426	2 747
BMP2	196	233
BTR70	196	233
T72	196	232
BRT60	195	256
2S1	274	299
BRDM2	274	298
D7	274	299
T62	273	299
ZIL131	274	299
ZSU23/4	274	299

数据集	Boost	SVM	FCN	LeNet	DCNN
17#	92.8	93.1	92.4	95.5	96.1
17#Ex	93.3	95.2	96.2	96.7	97.2

数据集	Boost	SVM	FCN	LeNet	DCNN
17#	92.8	93.1	92.4	95.5	96.1
17#Ex0.5	94.5	94.3	93.2	95.8	96.5
17#Ex1.0	93.3	95.2	96.2	96.7	97.2
17#Ex1.5	92.7	94.6	95.2	96.1	97.3
17#Ex2.0	93.1	95.0	94.9	95.6	95.9
17#Ex2.5	92.2	93.1	93.5	94.6	94.3

模型	Boost	SVM	FCN	LeNet	DCNN
wGAN-cp	89.5	88.7	89.3	91.1	91.1
wGAN-gp	91.2	90.5	90.2	94.8	95.2
LSGAN	92.9	93.3	94.7	95.4	95.6
Radar-GAN	93.3	95.2	96.2	96.7	97.2

[1]	仝卫国, 曾世超, 张立峰, 侯哲, 郭佳跃. 基于深度残差神经网络的电阻层析成像及流型辨识方法[J]. 系统仿真学报, 2022, 34(9): 2028-2036.
[2]	柴慧敏, 张勇, 李欣粤, 宋雅楠. 基于深度学习的空中目标威胁评估方法[J]. 系统仿真学报, 2022, 34(7): 1459-1467.
[3]	林硕, 安磊, 高治军, 单丹, 尚文利. 结合栈式自编码及长短时记忆的入侵检测研究[J]. 系统仿真学报, 2021, 33(6): 1288-1296.
[4]	程文聪, 史小康, 王志刚. 基于生成对抗网络的仿真卫星云图生成方法[J]. 系统仿真学报, 2021, 33(6): 1297-1306.
[5]	董书琴, 张斌. 面向不平衡数据的网络流量异常检测方法[J]. 系统仿真学报, 2021, 33(3): 679-689.
[6]	王霄汉, 张霖, 任磊, 谢堃钰, 王昆玉, 叶飞, 陈真. 基于强化学习的车间调度问题研究简述[J]. 系统仿真学报, 2021, 33(12): 2782-2791.
[7]	王步维, 王敏, 范谦, 王雅男, 章涵文, 乐云亮. 基于深度学习的晶体性质预测研究[J]. 系统仿真学报, 2021, 33(12): 2854-2863.
[8]	冯晓, 张辉, 周蕊, 乔璐, 魏东, 李丹丹, 张玉尧, 郑国清. 基于深度学习和籽粒双面特征的玉米品种识别[J]. 系统仿真学报, 2021, 33(12): 2983-2991.
[9]	杜金莲, 李淑飞, 金雪云. 三维烟雾流场超分辨率数据生成网络模型的研究[J]. 系统仿真学报, 2021, 33(10): 2381-2389.
[10]	刘瑞军, 王向上, 张晨, 章博华. 基于深度学习的视觉SLAM综述[J]. 系统仿真学报, 2020, 32(7): 1244-1256.
[11]	阴敬方, 朱登明, 石敏, 王兆其. 基于引导对抗网络的人体深度图像修补方法[J]. 系统仿真学报, 2020, 32(7): 1312-1321.
[12]	戢晓峰, 戈艺澄. 基于深度学习的节假日高速公路交通流预测方法[J]. 系统仿真学报, 2020, 32(6): 1164-1171.
[13]	孔锐, 谢玮, 雷泰. 基于神经网络的图像描述方法研究[J]. 系统仿真学报, 2020, 32(4): 601-611.
[14]	秦胜伟, 李重, 李金锋, 陈梓浩, 丁靖骞, 刘万顺. 校园漫游互动AR系统设计与实现[J]. 系统仿真学报, 2019, 31(7): 1367-1376.
[15]	叶继华, 时淑霞, 李汉曦, 王仕民, 杨思渝. 基于深度学习的驾驶关注区域检测研究与实现[J]. 系统仿真学报, 2019, 31(7): 1421-1428.