基于仿真SAR图像深度迁移学习的自动目标识别

doi:10.7523/j.issn.2095-6134.2020.04.011

中国科学院大学学报 ›› 2020, Vol. 37 ›› Issue (4): 516-524.DOI: 10.7523/j.issn.2095-6134.2020.04.011

基于仿真SAR图像深度迁移学习的自动目标识别

王泽隆^1,2, 徐向辉¹, 张雷³

1. 中国科学院电子学研究所, 北京 100190;
2. 中国科学院大学, 北京 100049;
3. 中国人民解放军95899部队, 北京 100076

收稿日期:2019-01-29 修回日期:2019-03-20 发布日期:2020-07-15
通讯作者: 王泽隆
基金资助:
国家重点研发计划（2017YFB0503001）资助

Study of deep transfer learning for SAR ATR based on simulated SAR images

WANG Zelong^1,2, XU Xianghui¹, ZHANG Lei³

1. Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China;
2. University of Chinese Academy of Sciences, Beijing 100049, China;
3. Unit 95899 of PLA, Beijing 100076, China

Received:2019-01-29 Revised:2019-03-20 Published:2020-07-15

摘要/Abstract

摘要： 使用深度卷积神经网络实现SAR图像的自动目标识别在训练过程中需要大量的标注数据。为解决由SAR实测数据获取成本高、标注数据量不足带来的问题，提出一种在由CReLU激活函数和批归一化改进的卷积神经网络上，使用仿真SAR图像提升最终目标识别性能的方法，把从大量仿真SAR图像学习到的有效知识迁移到实测SAR图像数据上。在训练中，先用仿真SAR图像预训练卷积神经网络，结合迁移学习的方法，有效地解决由SAR图像数据不足带来的过拟合问题。在MSTAR数据集上验证方法的有效性，识别准确率提高到99.78%，并在少量SAR图像样本数据上也取得不错的识别效果。

关键词: 合成孔径雷达, 仿真SAR图像, 迁移学习, 自动目标识别(ATR)

Abstract: Using deep convolutional neural networks to realize automatic target recognition of SAR requires a large amount of labeled data. In order to solve the problem caused by the scarcity of SAR real images, we propose a method for improving the target recognition performance of SAR by using simulated SAR images on convolutional neural networks improved by CReLU activation function and batch normalization. The method transfers the effective knowledge learned from a large number of simulated SAR images onto the real SAR images. In the training, the pre-trained convolutional neural networks can be obtained by training by using the simulated SAR images firstly, and the deep transfer learning method is used to effectively solve the problem caused by the insufficiency of SAR image data. The validation experiment is carried out on the MSTAR dataset. The highest recognition accuracy reaches 99.78%, and good recognition results are obtained based on a small amount of SAR image data.

Key words: SAR, simulated SAR images, transfer learning, ATR (automatic target recognition)

中图分类号:

TN957

王泽隆, 徐向辉, 张雷. 基于仿真SAR图像深度迁移学习的自动目标识别[J]. 中国科学院大学学报, 2020, 37(4): 516-524.

WANG Zelong, XU Xianghui, ZHANG Lei. Study of deep transfer learning for SAR ATR based on simulated SAR images[J]. , 2020, 37(4): 516-524.

参考文献

[1] Cummimg I, Wong F. 合成孔径雷达成像:算法与实现[M]. 北京:电子工业出版社, 2012:2-11.
[2] Mcnairn H, Brisco B. The application of C-band polarimetric SAR for agriculture:a review[J]. Canadian Journal of Remote Sensing, 2004, 30(3):525-542.
[3] Ross T D, Worrell S W, Velten V J, et al. Standard SAR ATR evaluation experiments using the MSTAR public release data set[C]//Algorithms for Synthetic Aperture Radar Imagery V. Orlando, United States:International Society for Optics and Photonics, 1998, 3370:566-574.
[4] Anagnostopoulos G C. SVM-based target recognition from synthetic aperture radar images using target region outline descriptors[J]. Nonlinear Analysis:Theory, Methods & Applications, 2009, 71(12):e2934-e2939.
[5] Song S, Xu B, Yang J. SAR target recognition via supervised discriminative dictionary learning and sparse representation of the SAR-HOG feature[J]. Remote Sensing, 2016, 8(8):683.
[6] Park J I, Kim K T. Modified polar mapping classifier for SAR automatic target recognition[J]. IEEE Transactions on Aerospace and Electronic Systems, 2014, 50(2):1092-1107.
[7] Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks[C]//Advances in neural information processing systems. Nevada, USA:Curran Associates Inc, 2012:1097-1105.
[8] Schmidhuber J. Deep learning in neural networks:an overview[J]. Neural Networks, 2015, 61:85-117.
[9] Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions[C]//IEEE Conference on Computer Vision and Pattern Recognition. Boston:IEEE Computer Society, 2015:1-9.
[10] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas:IEEE Computer Society, 2016:770-778.
[11] Malmgren-Hansen D, Kusk A, Dall J, et al. Improving SAR automatic target recognition models with transfer learning from simulated data[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(9):1484-1488.
[12] Morgan D A E. Deep convolutional neural networks for ATR from SAR imagery[C]//Algorithms for Synthetic Aperture Radar Imagery XXII. Maryland, United States:International Society for Optics and Photonics, 2015, 9475:94750F.
[13] 李松, 魏中浩, 张冰尘, 等. 深度卷积神经网络在迁移学习模式下的SAR目标识别[J]. 中国科学院大学学报, 2018, 35(1):75-83.
[14] Huang Z, Pan Z, Lei B. Transfer learning with deep convolutional neural network for SAR target classification with limited labeled data[J]. Remote Sensing, 2017, 9(9):907.
[15] 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.
[16] 董纯柱, 胡利平, 朱国庆,等. 地面车辆目标高质量SAR图像快速仿真方法[J]. 雷达学报, 2015, 4(3):351-360.
[17] Kusk A, Abulaitijiang A, Dall J. Synthetic SAR image generation using sensor, terrain and target models[C]//Proceedings of EUSAR 2016:11th European Conference on Synthetic Aperture Radar. Hamburg, Germany:VDE, 2016:1-5.
[18] Shang W, Sohn K, Almeida D, et al. Understanding and improving convolutional neural networks via concatenated rectified linear units[C]//International Conference on Machine Learning. New York:ICML, 2016:2217-2225.
[19] Ioffe S. Batch renormalization:Towards reducing minibatch dependence in batch-normalized models[C]//Advances in Neural Information Processing Systems. Long Beach:NIPS, 2017:1945-1953.
[20] Glorot X, Bengio Y. Understanding the difficulty of training deep feedforward neural networks[C]//Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics. Sardinia:AISTATS, 2010:249-256.
[21] He K, Zhang X, Ren S, et al. Delving deep into rectifiers:Surpassing human-level performance on imagenet classification[C]//Proceedings of the IEEE International Conference on Computer Vision. Washington, DC, USA:IEEE Computer Society, 2015:1026-1034.
[22] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[C]//Advances in Neural Information Processing Systems. Long Beach, NIPS, 2017:5998-6008.

基于仿真SAR图像深度迁移学习的自动目标识别

Study of deep transfer learning for SAR ATR based on simulated SAR images

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

访问统计

联系我们

[1]	尹迪, 韩冰, 孙吉利, 赵良波, 洪文, 胡玉新. 天线扫描对TOPSAR辐射影响的校正方法[J]. 中国科学院大学学报, 2021, 38(6): 791-799.
[2]	林世明, 张毅, 王吉利, 赵爽. 星载双基InSAR系统中的电离层色散效应分析与校正[J]. 中国科学院大学学报, 2021, 38(6): 809-816.
[3]	史达亨, 刘立刚, 周斌, 卜智勇. 跨时间迁移的多源无线信号指纹融合定位方法[J]. 中国科学院大学学报, 2021, 38(6): 817-824.
[4]	刘雨洁, 齐向阳. 基于长时间间隔序贯SAR图像的运动舰船跟踪方法[J]. 中国科学院大学学报, 2021, 38(5): 642-648.
[5]	潘志刚, 王超, 曹舸. 基于距离方程组的机载SAR图像立体定位方法[J]. 中国科学院大学学报, 2020, 37(6): 798-804.
[6]	田东, 禹卫东. 电离层色散效应对星载SAR成像的影响[J]. 中国科学院大学学报, 2019, 36(3): 376-384.
[7]	王健飞, 张卫强, 刘加. 基于多状态跳转模型的场景独立音频事件检测方法[J]. 中国科学院大学学报, 2019, 36(2): 218-225.
[8]	邹浩, 林赟, 洪文. 基于多角度合成SAR图像的目标识别性能分析[J]. 中国科学院大学学报, 2019, 36(2): 226-234.
[9]	蒋莎, 仇晓兰, 韩冰, 胡文龙, 卢晓军. 一种基于普通分布目标的极化SAR幅相不平衡快速评价方法[J]. 中国科学院大学学报, 2019, 36(2): 280-288.
[10]	李松, 魏中浩, 张冰尘, 洪文. 深度卷积神经网络在迁移学习模式下的SAR目标识别[J]. 中国科学院大学学报, 2018, 35(1): 75-83.
[11]	肖垚, 刘畅. 基于稀疏求解的改进PCA方法在SAR目标识别中的应用[J]. 中国科学院大学学报, 2018, 35(1): 84-88.
[12]	杨国铮, 禹晶, 孙卫东. 基于相对全变分的复杂背景SAR图像舰船尾迹检测[J]. 中国科学院大学学报, 2017, 34(6): 734-742.
[13]	陈立福, 王思雨, 袁志辉, 王静, 邢学敏, 武鸿. 非线性ECS自配准成像算法中相位保持精度分析[J]. 中国科学院大学学报, 2017, 34(5): 640-646.
[14]	潘志刚, 王小龙, 李志勇. SAR原始数据压缩的自适应比特分配BAQ算法[J]. 中国科学院大学学报, 2017, 34(1): 106-111.
[15]	杨洋, 王岩飞. 一种基于混合模型的合成孔径雷达自聚焦算法[J]. 中国科学院大学学报, 2016, 33(5): 656-663.