[1]曹磊,王强,史润佳,等.基于改进RPN的Faster-RCNN网络SAR图像车辆目标检测方法[J].东南大学学报(自然科学版),2021,51(1):87-91.[doi:10.3969/j.issn.1001-0505.2021.01.012]
 Cao Lei,Wang Qiang,Shi Runjia,et al.Method for vehicle target detection on SAR image based on improved RPN in Faster-RCNN[J].Journal of Southeast University (Natural Science Edition),2021,51(1):87-91.[doi:10.3969/j.issn.1001-0505.2021.01.012]
点击复制

基于改进RPN的Faster-RCNN网络SAR图像车辆目标检测方法()
分享到:

《东南大学学报(自然科学版)》[ISSN:1001-0505/CN:32-1178/N]

卷:
51
期数:
2021年第1期
页码:
87-91
栏目:
电磁场与微波技术
出版日期:
2021-01-20

文章信息/Info

Title:
Method for vehicle target detection on SAR image based on improved RPN in Faster-RCNN
作者:
曹磊王强史润佳蒋忠进
东南大学毫米波国家重点实验室, 南京 210096
Author(s):
Cao Lei Wang Qiang Shi Runjia Jiang Zhongjin
State Key Laboratory of Millimeter Waves, Southeast University, Nanjing 210096, China
关键词:
SAR图像 车辆目标检测 卷积神经网络 Faster-RCNN 候选区域生成网络
Keywords:
SAR images vehicle target detection convolutional neural network Faster-RCNN region proposal network(RPN)
分类号:
TN957.51
DOI:
10.3969/j.issn.1001-0505.2021.01.012
摘要:
针对传统Faster-RCNN方法中候选区域生成网络(RPN)模块在进行目标检测时对目标特征提取不够充分的问题,提出一种基于改进RPN的Faster-RCNN网络SAR图像车辆目标检测方法.首先基于VGG-16网络提取出图片的多层特征,然后利用卷积核对最深的3个特征层作进一步的特征提取和正则化处理,最后对处理后的3个特征层进行信息融合.利用MSTAR数据集中车辆目标SAR图像和自然背景SAR图像,通过图像分割和贴图的方式制作了SAR场景数据集,对所改进网络进行训练和测试.实验结果表明,在SAR图像车辆目标检测中,与传统RPN相比,改进RPN收敛速度更快,不仅将检测结果的查准率从97.7%提高到了99.7%,虚警率明显降低,而且泛化性能更强,针对训练范围以外的目标,能将查准率由98.0%提高到99.0%.
Abstract:
Aiming at the problem that the region proposal network(RPN)module couldn’t adequately extract target features when performing target detection in the traditional Faster-RCNN method, a method for vehicle target detection on the SAR image based on the improved RPN in Faster-RCNN was proposed. First, the multi-layer features of the image were extracted based on the VGG-16 network. Then the deepest three feature layers were further extracted and regularized using convolution kernels. Finally, the information fusion was performed on the three processed feature layers. Using the vehicle target SAR image and the natural background SAR image in the MSTAR data set, the SAR scene data set was created by image segmentation and texture, and the improved network was trained and tested. Experimental results show that in the detection of the vehicle targets on SAR images, compared with the traditional RPN, the improved RPN has a faster convergence speed, improving the accuracy of the test results from 97.7% to 99.7%, with a lower false alarm rate, and has stronger generalization performance. For targets outside the training range, the accuracy can be increased from 98.0% to 99.0%.

参考文献/References:

[1] 郑远攀, 李广阳, 李晔. 深度学习在图像识别中的应用研究综述[J]. 计算机工程与应用, 2019, 55(12): 20-36.
  Zheng Y P, Li G Y, Li Y. Survey of application of deep learning in image recognition[J]. Computer Engineering and Applications, 2019, 55(12): 20-36.(in Chinese)
[2] Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60(6): 84-90. DOI: 10.1145/3065386.
[3] 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.
  Xu F, Wang H P, Jin Y Q. Deep learning as applied in SAR target recognition and terrain classification[J]. Journal of Radars, 2017, 6(2): 136-148.(in Chinese)
[4] 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.
  Tian Z Z, Zhan R H, Hu J M, et al. SAR ATR based on convolutional neural network[J]. Journal of Radars, 2016, 5(3): 320-325.(in Chinese)
[5]邹浩, 林赟, 洪文. 采用深度学习的多方位角SAR图像目标识别研究[J]. 信号处理, 2018, 34(5): 513-522.
  Zou H, Lin Y, Hong W. Research on multi-aspect SAR images target recognition using deep learning[J]. Journal of Signal Processing, 2018, 34(5): 513-522.(in Chinese)
[6] 薛媛. 基于深度神经网络的SAR自动目标识别方法研究[D]. 成都: 电子科技大学, 2019.
  Xue Y. Research on SAR automatic target recognition based on deep neural network[D]. Chengdu: University of Electronic Science and Technology of China, 2019.(in Chinese)
[7] 刘彬. 基于卷积神经网络的SAR图像目标检测及分类方法研究[D]. 西安: 西安电子科技大学, 2017.
  Liu B. Target detection and classification methods based on convolutional neural network for SAR image[D]. Xi’an: Xidian University, 2017.(in Chinese)
[8] 张笑. 基于深度学习的SAR图像目标识别算法研究[D].南京:南京航空航天大学,2018.
  Zhang X. Research on SAR image target recognition based on deep learning[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2018.(in Chinese)
[9] Shang R H, Wang J M, Jiao L C, et al. SAR targets classification based on deep memory convolution neural networks and transfer parameters[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2018, 11(8): 2834-2846. DOI:10.1109/jstars.2018.2836909.
[10] Girshick R,Donahue J, Darrell T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]// 27th IEEE Conference on Computer Vision and Pattern Recognition. Columbus, OH, USA, 2014:580-587. DOI: 10.1109/CVPR.2014.81.
[11] He K M, Zhang X Y, Ren S Q, et al. Spatial pyramid pooling in deep convolutional networks for visual recognition[J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9): 1904-1916. DOI:10.1109/tpami.2015.2389824.
[12] Girshick R. Fast R-CNN[C]// Proceedings of the IEEE Conference on Computer Vision. Santiago, Chile, 2015:1440-1448.
[13] Ren S Q, He K M, Girshick R, et al. Faster R-CNN: Towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149. DOI:10.1109/tpami.2016.2577031.
[14] 杜兰, 刘彬, 王燕, 等. 基于卷积神经网络的SAR图像目标检测算法[J]. 电子与信息学报, 2016, 38(12): 3018-3025.
  Du L, Liu B, Wang Y, et al. Target detection method based on convolutional neural network for SAR image[J]. Journal of Electronics & Information Technology, 2016, 38(12): 3018-3025.(in Chinese)
[15] 王思雨, 高鑫, 孙皓, 等. 基于卷积神经网络的高分辨率SAR图像飞机目标检测方法[J]. 雷达学报, 2017, 6(2): 195-203.
  Wang S Y, Gao X, Sun H, et al. An aircraft detection method based on convolutional neural networks in high-resolution SAR images[J]. Journal of Radars, 2017, 6(2): 195-203.(in Chinese)
[16] 常沛, 夏勇, 李玉景, 等. 基于CNN的SAR车辆目标检测[J]. 雷达科学与技术, 2019, 17(2): 220-224, 231.
  Chang P, Xia Y, Li Y J, et al. SAR vehicle target detection based on CNN[J]. Radar Science and Technology, 2019, 17(2): 220-224, 231.(in Chinese)

相似文献/References:

[1]曹兰英,夏良正,张昆辉.基于小波域MRF模型的SAR图像分割[J].东南大学学报(自然科学版),2004,34(6):847.[doi:10.3969/j.issn.1001-0505.2004.06.029]
 Cao Lanying,Xia Liangzheng,Zhang Kunhui.SAR image segmentation using MRF model in wavelet domain[J].Journal of Southeast University (Natural Science Edition),2004,34(1):847.[doi:10.3969/j.issn.1001-0505.2004.06.029]

备注/Memo

备注/Memo:
收稿日期: 2020-06-12.
作者简介: 曹磊(1994—),男,硕士生;蒋忠进(联系人),男,博士,副教授,zjjiang@seu.edu.cn.
引用本文: 曹磊,王强,史润佳,等.基于改进RPN的Faster-RCNN网络SAR图像车辆目标检测方法[J].东南大学学报(自然科学版),2021,51(1):87-91. DOI:10.3969/j.issn.1001-0505.2021.01.012.
更新日期/Last Update: 2021-01-20