吉林大学学报(工学版) ›› 2020, Vol. 50 ›› Issue (5): 1778-1787.doi: 10.13229/j.cnki.jdxbgxb20190490

• 计算机科学与技术 • 上一篇    

基于深度学习的SAR图像道路识别新方法

谌华1,2,3(),郭伟1,2(),闫敬文4,卓文浩4,吴良斌5   

  1. 1.中国科学院 微波遥感技术重点实验室，北京 0090
    2.中国科学院 国家空间科学中心，北京 100190
    3.中国科学院大学 研究生院，北京 100049
    4.汕头大学 工学院，广东 汕头 515063
    5.中航雷达与电子设备研究院，江苏 无锡 214063
  • 收稿日期:2019-05-20 出版日期:2020-09-01 发布日期:2020-09-16
  • 通讯作者: 郭伟 E-mail:e_chenhua@163.com;guowei@mirslab.cn
  • 作者简介:谌华（1979-）,男,高级工程师,博士研究生.研究方向：SAR图像处理，机器学习.E-mail:e_chenhua@163.com
  • 基金资助:
    国家自然科学基金项目(61672335);广东省创新强校基金项目(2016KZDXM012)

A new deep learning method for roads recognition from SAR images

Hua CHEN1,2,3(),Wei GUO1,2(),Jing-wen YAN4,Wen-hao ZHUO4,Liang-bin WU5   

  1. 1.Key Laboratory of Microwave Remote Sensing, Chinese Academy of Sciences, Beijing 0090, China
    2.National Space Science Center, Chinese Academy of Sciences, Beijing 100190, China
    3.Graduate School of University of Chinese Academy of Sciences, Beijing 100049, China
    4.College of Engineering, Shantou University, Shantou 515063, China
    5.Radar and Electronic Equipment Research Institute of China Aviation, Wuxi 214063, China
  • Received:2019-05-20 Online:2020-09-01 Published:2020-09-16
  • Contact: Wei GUO E-mail:e_chenhua@163.com;guowei@mirslab.cn

RICH HTML

 31   

PDF (PC)

485

摘要：

针对传统合成孔径雷达(SAR)图像道路识别步骤繁杂的问题，提出了一种新的基于深度学习的SAR图像道路识别方法。首先，在原有全卷积神经网络(FCN)的基础上通过改进激活函数构造一种新的卷积神经网络M-FCN，有效缓解了道路信息丢失问题；然后，将该卷积神经网络和自主构建的道路标签集应用于模拟SAR和真实SAR图像道路识别实验中，提高了鲁棒性。实验结果表明：与支持向量机(SVM)、传统全卷积神经网络和其他算法比较，该算法可以用来识别道路特征，并具有较高的精度和可靠性。

关键词: 信息处理技术, 合成孔径雷达图像, 道路识别, 深度学习, 全卷积神经网络, 最大特征映射

Abstract:

Deep learning is an effective technical method to enhance the recognition accuracy of remote sensing image target. To solve the problem of the complicated steps in road recognition from Synthetic Aperture Radar(SAR) images, this paper proposes a SAR image road recognition method based on deep learning. First, based on the traditional Full Convolution Neural Network(FCNN), a New Convolution Neural Network(NCNN) is constructed by revising the activation function to effectively alleviate the loss of road information. Then, the NCNN is applied to road recognition experiments of simulated SAR images and real SAR images with self-constructed road label sets to improve the robustness of the proposed method. The experimental results show that the NCNN can be used to identify the overall characteristics of the road with higher accuracy and better reliability in comparison with the Support Vector Machine(SVM), traditional FCNN and other algorithms.

Key words: information processing technology, synthetic aperture radar image, roads recognition, deep learning, fully convolutional network, max feature map

中图分类号: 

  • TP391.4

图1

典型的全卷积神经网络模型框架图"

图2

基于MFM的卷积神经网络框架图"

图3

M-FCN框架的网络层算法流程图"

图4

模拟的SAR实验数据"

图5

各种算法道路识别的结果对比"

表1

模拟SAR图像不同道路识别方法结果对比 (%)"

评价指标SVMFCNFCN-ReLUM-FCN
IoU41.760.567.669.4
ACC83.595.196.096.2

图6

光学道路影像标签及模拟的SAR图像"

图7

SNR变化对应M-FCN识别SAR道路特征准确率的变化"

表2

SNR变化对M-FCN提取道路特征准确率的影响 (%)"

SNRmIoUACC
0.259.095.4
0.362.195.7
0.463.795.7
0.563.995.8
0.664.795.9
0.765.295.8
0.866.296.0
0.969.496.3

图8

各种算法在真实SAR图像中道路识别结果的对比"

1 Bajcsy R, Tavakoli M. Computer recognition of roads from satellite pictures[J]. IEEE Transactions on Systems Man and Cybernetics, 1976, 6(9): 623-637.
2 张永宏, 何静, 阚希, 等. 遥感图像道路提取方法综述[J]. 计算机工程与应用, 2018, 54(13): 1-10.
Zhang Yong-hong, He Jin, Kan Xi, et al. Summary of road extraction methods for remote sensing images[J]. Computer Engineering and Applications, 2018, 54(13): 1-10.
3 Samadani R, Vesecky J F. Finding curvilinear features in speckled images[J]. IEEE Transactions on Geoscience and Remote Sensing, 1990, 28(4): 669-673.
4 Tupin F, Maitre H, Mangin J F, et al. Detection of linear features in SAR images: application to road network extraction[J]. IEEE Transactions on Geoscience and Remote Sensing, 1998, 36(2): 434-453.
5 Negri M, Gamba P, Lisini G, et al. Junction-aware extraction and regularization of urban road networks in high-resolution SAR images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2006, 44(10): 2962-2971.
6 Lisini G, Tison C, Tupin F, et al. Feature fusion to improve road network extraction in high-resolution SAR images[J]. IEEE Geoscience and Remote Sensing Letters, 2006, 3(2): 217-221.
7 Lu P P, Du K N, Yu W D, et al. A new region growing-based method for road network extraction and its application on different resolution SAR images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(12): 4772-4783.
8 Jin R, Zhou W, Yin J, et al. CFAR line detector for polarimetric SAR images using wilks’ test statistic[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(5): 711-715.
9 Xiao F H, Chen Y, Tong L, et al. Road detection in high-resolution SAR images using duda and path operators[C]∥IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Beijing, 2016: 1266-1269.
10 Wei Q R, Feng D Z. Extraction line feature in SAR images through image edge fields[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(4): 540-544.
11 程江华, 高贵, 库锡树, 等. SAR 图像道路网提取方法综述[J]. 中国图象图形学报, 2013, 18(1): 11-23．
Cheng Jiang-hua, Gao Gui, Xi-shu Ku, et al. Review of road network extraction from SAR images[J]. Journal of Image and Graphics, 2013, 18(1): 11-23.
12 Tupin F, Houshmand B, Datcu M. Road detection in dense urban areas using SAR imagery and the usefulness of multiple views[J]. IEEE Transactions on Geoscience and Remote Sensing, 2002, 40(11): 2405-2414.
13 Jeon B K, Jang J H, Hong K S. Road detection in spaceborne SAR images using a genetic algorithm[J]. IEEE Transactions on Geoscience and Remote Sensing, 2002, 40(1): 22-29.
14 张广伟, 张永红. 基于链码优化的SAR 影像城市道路网络提取[J]. 遥感学报, 2008, 12(4): 620-625.
Zhang Guang-wei, Zhang Yong-hong. Road network extraction of urban areasin SAR image based on optimization of chain code[J]. Journal of Remote Sensing, 2008, 12(4): 620-625.
15 洪日昌, 吴秀清, 刘媛, 等. 低分辨率遥感影像中道路的全自动提取方法研究[J]. 遥感学报, 2008, 12(1): 36-45.
Hong Ri-chang, Wu Xiu-qing, Liu Yuan, et al. Research on roads automatic extraction from low resolution remote sensing image[J]. Journal of Remote Sensing, 2008, 12(1): 36-45.
16 Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks[C]∥International Conference on Neural Information Processing Systems, Curran Associates Inc., 2012: 1097-1105.
17 Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition[J]. Computer Vision and Pattern Recognition, ArXiv: 1409.1556, 2014.
18 Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions[C]∥IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, USA, 2015: 1-9.
19 He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]∥IEEE Conference on Computer Vision and Pattern Recognition, LasVegas, 2016: 770-778.
20 Chen S Z, Wang H P, Xu F, et al. Target classification using the deep convolutional networks for SAR images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(8): 4806-4817.
21 Wang Y, Wang C, Zhang H. Combining a single shot multibox detector with transfer learning for ship detection using sentinel-1 SAR images[J]. Remote Sensing Letters, 2018, 9(8): 780-788.
22 Wang Y Y, Wang C, Zhang H, et al. Automatic ship detection based on retina net using multi-resolution Gaofen-3 imagery[J]. Remote Sensing, 2019, 11(5): 531.
23 Corentin H, Majid A S, Nina M. Road segmentation in SAR satellite images with deep fully convolutional neural networks[J]. IEEE Geoscience and Remote Sensing Letters, 2018, 15(12): 1867-1871.
24 Jonathan L, Evan S, Trevor D. Fully convolutional networks for semantic segmentation[C]∥IEEE Conference on Computer Vision and Pattern Recognition, Boston, 2015: 3431-3440.
25 Wu X, He R, Sun Z. A lightened CNN for deep face representation[J]. Computer Science,arXiv:1511.02683, 2015.
26 Cheng G, Wang Y, Xu S, et al. Automatic road detection and centerline extraction via cascaded end-to-end convolutional neural network[J]. IEEE Transactions on Geoscience & Remote Sensing, 2017, 55(6): 3322-3337.
[1] 史再峰,李金卓,曹清洁,李慧龙,胡起星. 基于生成对抗网络的低剂量能谱层析成像去噪算法[J]. 吉林大学学报(工学版), 2020, 50(5): 1755-1764.
[2] 王柯俨,王迪,赵熹,陈静怡,李云松. 基于卷积神经网络的联合估计图像去雾算法[J]. 吉林大学学报(工学版), 2020, 50(5): 1771-1777.
[3] 赵宏伟,刘晓涵,张媛,范丽丽,龙曼丽,臧雪柏. 基于关键点注意力和通道注意力的服装分类算法[J]. 吉林大学学报(工学版), 2020, 50(5): 1765-1770.
[4] 程艳芬,姚丽娟,袁巧,陈先桥. 水下视频图像清晰化方法[J]. 吉林大学学报(工学版), 2020, 50(2): 668-677.
[5] 郜峰利,陶敏,李雪妍,何昕,杨帆,王卓,宋俊峰,佟丹. 基于深度学习的CT影像脑卒中精准分割[J]. 吉林大学学报(工学版), 2020, 50(2): 678-684.
[6] 张薇,韩勇,金铭,乔晓林. 基于托普利兹矩阵集重构的相干信源波达方向估计[J]. 吉林大学学报(工学版), 2020, 50(2): 703-710.
[7] 刘富, 权美静, 王柯, 刘云, 康冰, 韩志武, 侯涛. 仿蝎子振源定位机理的位置指纹室内定位方法[J]. 吉林大学学报(工学版), 2019, 49(6): 2076-2082.
[8] 于晓辉,张志成,李新波,孙晓东. 基于状态空间模型的指数衰减正弦信号参数估计[J]. 吉林大学学报(工学版), 2019, 49(6): 2083-2088.
[9] 马子骥,卢浩,董艳茹. 双通道单图像超分辨率卷积神经网络[J]. 吉林大学学报(工学版), 2019, 49(6): 2089-2097.
[10] 郭继昌,吴洁,郭春乐,朱明辉. 基于残差连接卷积神经网络的图像超分辨率重构[J]. 吉林大学学报(工学版), 2019, 49(5): 1726-1734.
[11] 曹运合,曾丽,王宇. 基于特征空间的子阵级自适应和差波束测角方法[J]. 吉林大学学报(工学版), 2019, 49(5): 1735-1744.
[12] 徐谦,李颖,王刚. 基于深度学习的行人和车辆检测[J]. 吉林大学学报(工学版), 2019, 49(5): 1661-1667.
[13] 卢洋,王世刚,赵文婷,赵岩. 基于离散Shearlet类别可分性测度的人脸表情识别方法[J]. 吉林大学学报(工学版), 2019, 49(5): 1715-1725.
[14] 董超,刘晶红,徐芳,王仁浩. 光学遥感图像舰船目标快速检测方法[J]. 吉林大学学报(工学版), 2019, 49(4): 1369-1376.
[15] 王柯俨,胡妍,王怀,李云松. 结合天空分割和超像素级暗通道的图像去雾算法[J]. 吉林大学学报(工学版), 2019, 49(4): 1377-1384.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
版权所有 © 《吉林大学学报(工学版)》编辑部
地址：长春市人民大街5988号　电话：+86-431-85095297　传真：+86-431-85094074　E-mail: xbgxb@jlu.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发