基于深度学习的探地雷达二维剖面图像结构特征检测方法

王辉; 欧阳缮; 刘庆华; 廖可非; 周丽军

doi:10.11999/JEIT211032

基于深度学习的探地雷达二维剖面图像结构特征检测方法

doi: 10.11999/JEIT211032

王辉^{1, 2, 3},
欧阳缮^{1, 3, ,},
刘庆华¹,
廖可非¹,
周丽军⁴

1.
桂林电子科技大学信息与通信学院桂林 541004
2.
贺州学院人工智能学院贺州 542899
3.
桂林电子科技大学卫星导航定位与位置服务国家地方联合工程研究中心桂林 541004
4.
山西省交通科技研发有限公司太原 030032

基金项目: 国家自然科学基金(61871425, 61861011, 61631019)，广西创新驱动发展专项(桂科AA21077008)，山西省交通运输厅科技项目(2019-1-18)

详细信息

作者简介:
王辉：男，1982年生，博士生，副教授，研究方向为探地雷达信号处理、深度学习

欧阳缮：男，1960年生，博士，教授，博士生导师，研究方向为智能信息处理

刘庆华：女，1973年生，博士，教授，硕士生导师，研究方向为智能信息处理、深度学习

廖可非：男，1984年生，博士，副教授，硕士生导师，研究方向为探地雷达信号处理

周丽军：女，1984年生，博士，高级工程师，研究方向为探地雷达信号处理

通讯作者:
欧阳缮　hmoysh@guet.edu.cn

中图分类号: TN957.52
计量
- 文章访问数: 975
- HTML全文浏览量: 1232
- PDF下载量: 175
- 被引次数: 0
出版历程
- 收稿日期: 2021-09-27
- 修回日期: 2022-02-18
- 录用日期: 2022-02-23
- 网络出版日期: 2022-03-07
- 刊出日期: 2022-04-18

Structure Feature Detection Method for Ground Penetrating Radar Two-Dimensional Profile Image Based on Deep Learning

WANG Hui^{1, 2, 3},
OUYANG Shan^{1, 3
, ,},
LIU Qinghua¹,
LIAO Kefei¹,
ZHOU Lijun⁴

1.
School of Information and Communication, Guilin University of Electronic Technology, Guilin 541004, China
2.
School of Artificial Intelligence, Hezhou University, Hezhou 542899, China
3.
Satellite Navigation Positioning and Location Service National and Local Joint Engineering Research Center, Guilin University of Electronic Technology, Guilin 541004, China
4.
Shanxi Transportation Technology R&D Co., LTD, Taiyuan 030032, China

Funds: The National Natural Science Foundation of China (61871425, 61861011, 61631019), Guangxi Special Fund Project for Innovation-driven Development (GuikeAA21077008), Shanxi Transportation Department Projects (2019-1-18)

摘要

摘要: 该文针对探地雷达(GPR) 2维剖面图像中目标特征提取困难及其识别精度较低等问题，采用深度学习方法来提取2维剖面图像中目标的特征双曲线。根据GPR工作的物理机制，设计了一种级联结构的卷积神经网络(CNN)，先检测并去除回波数据中的直达波干扰信号，再利用CNN得到B扫描(B-SCAN)图像的特征图，并对特征信号进行分类识别以提取目标的特征双曲线。同时，为处理各种干扰信号影响目标特征双曲线结构完整性的问题，提出了一种基于方向引导的特征数据补全方法，提高了目标特征双曲线识别的准确率。与方向梯度直方图(HOG)算法、单级式目标检测(YOLOV3)算法和更快速的区域卷积神经网络(Faster RCNN)算法相比，在综合评价指标F上该文方法的检测结果是最优的。
- 探地雷达 /
- 深度学习 /
- 特征检测 /
- 卷积神经网络
Abstract: To solve the problem of the difficulty of target feature extraction and low recognition accuracy in Ground Penetrating Radar (GPR) two-dimensional profile, a deep learning method is used to extract the characteristic hyperbola of targets in B-SCAN image. Physics-based of GPR, a cascade Convolutional Neural Network (CNN) is designed to detect and remove the direct wave interference signal in the echo data. Then, the B-SCAN image is obtained by CNN, and the characteristic signals are classified and recognized to extract the characteristic hyperbola of the target. Meanwhile, in order to deal with the problem that the interference signals affect the structural integrity of the feature hyperbola, a feature data completion method based on directional guidance is proposed to improve the accuracy of the feature hyperbola recognition results. Compared with Histogram of Oriented Gradients(HOG) algorithm, You Only Look Once V3(YOLOV3) algorithm and Faster Region-based Convolutional Neural Network(Faster RCNN) algorithm, the detection result of the proposed method is the best in the comprehensive evaluation index F.
- Ground Penetrating Radar(GPR) /
- Deep learning /
- Feature detection /
- Convolutional Neural Network(CNN)

HTML全文

图 1 目标特征检测框架

下载: 全尺寸图片幻灯片

图 2 GPR反射测量模式及仿真2维剖面图像

下载: 全尺寸图片幻灯片

图 3 B-SCAN图像特征提取的级联CNN结构

下载: 全尺寸图片幻灯片

图 4 目标特征双曲线提取网络训练模型的损失曲线

下载: 全尺寸图片幻灯片

图 5 结构信息不完整的GPR B-SCAN特征双曲线示意图

下载: 全尺寸图片幻灯片

图 6 提取GPR B-SCAN图像的特征图

下载: 全尺寸图片幻灯片

图 7 目标双曲线特征数据补全

下载: 全尺寸图片幻灯片

图 8 B-SCAN图像中的目标特征双曲线检测结果对比

下载: 全尺寸图片幻灯片

图 9 不同算法的F均值统计结果

下载: 全尺寸图片幻灯片

参考文献(24)

[1]	PAMBUDI A D, FAUß M, AHMAD F, et al. Minimax robust landmine detection using forward-looking ground-penetrating radar[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020, 58(7): 5032–5041. doi: 10.1109/TGRS.2020.2971956
[2]	EBRAHIM S M, MEDHAT N I, MANSOUR K K, et al. Examination of soil effect upon GPR detectability of landmine with different orientations[J]. NRIAG Journal of Astronomy and Geophysics, 2018, 7(1): 90–98. doi: 10.1016/j.nrjag.2017.12.004
[3]	郝彤, 赵杰. 面向双曲线形态的探地雷达图像识别技术综述[J]. 电子学报, 2019, 47(6): 1366–1372. doi: 10.3969/j.issn.0372-2112.2019.06.025 HAO Tong and ZHAO Jie. A brief review of the hyperbola signature recognition techniques for ground penetrating radar[J]. Acta Electronica Sinica, 2019, 47(6): 1366–1372. doi: 10.3969/j.issn.0372-2112.2019.06.025
[4]	侯斐斐, 施荣华, 雷文太, 等. 面向探地雷达B-scan图像的目标检测算法综述[J]. 电子与信息学报, 2020, 42(1): 191–200. doi: 10.11999/JEIT190680 HOU Feifei, SHI Ronghua, LEI Wentai, et al. A review of target detection algorithm for GPR B-scan processing[J]. Journal of Electronics &Information Technology, 2020, 42(1): 191–200. doi: 10.11999/JEIT190680
[5]	LUO T X H, LAI W W L, CHANG R K W, et al. GPR imaging criteria[J]. Journal of Applied Geophysics, 2019, 165: 37–48. doi: 10.1016/j.jappgeo.2019.04.008
[6]	YANG Tianchun, DAI Shixin, DENG Hanyang, et al. Application research of GPR in quality detection of urban underground jacked pipes construction[J]. IOP Conference Series:Earth and Environmental Science, 2019, 295(4): 042069. doi: 10.1088/1755-1315/295/4/042069
[7]	FIROOZABADI R, MILLER E L, RAPPAPORT C M, et al. Subsurface sensing of buried objects under a randomly rough surface using scattered electromagnetic field data[J]. IEEE Transactions on Geoscience and Remote Sensing, 2007, 45(1): 104–117. doi: 10.1109/TGRS.2006.883462
[8]	MINET J, LAMBOT S, SLOB E C, et al. Soil surface water content estimation by full-waveform GPR signal inversion in the presence of thin layers[J]. IEEE Transactions on Geoscience and Remote Sensing, 2010, 48(3): 1138–1150. doi: 10.1109/TGRS.2009.2031907
[9]	CAPINERI L, GRANDE P, and TEMPLE J A G. Advanced image-processing technique for real-time interpretation of ground-penetrating radar images[J]. International Journal of Imaging Systems and Technology, 1998, 9(1): 51–59. doi: 10.1002/(SICI)1098-1098(1998)9:1<51::AID-IMA7>3.0.CO;2-Q
[10]	MAAS C and SCHMALZL J. Using pattern recognition to automatically localize reflection hyperbolas in data from ground penetrating radar[J]. Computers & Geosciences, 2013, 58: 116–125. doi: 10.1016/j.cageo.2013.04.012
[11]	TORRIONE P A, MORTON K D, SAKAGUCHI R, et al. Histograms of oriented gradients for landmine detection in ground-penetrating radar data[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014, 52(3): 1539–1550. doi: 10.1109/TGRS.2013.2252016
[12]	FRIGUI H and GADER P. Detection and discrimination of land mines in ground-penetrating radar based on edge histogram descriptors and a possibilistic k-nearest neighbor classifier[J]. IEEE Transactions on Fuzzy Systems, 2009, 17(1): 185–199. doi: 10.1109/TFUZZ.2008.2005249
[13]	PHAM M T and LEFÈVRE S. Buried object detection from B-scan ground penetrating radar data using Faster-RCNN[C]. IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium, Valencia, Spain, 2018: 6804–6807.
[14]	LEI Wentai, HOU Feifei, XI Jingchun, et al. Automatic hyperbola detection and fitting in GPR B-scan image[J]. Automation in Construction, 2019, 106: 102839. doi: 10.1016/j.autcon.2019.102839
[15]	王辉, 欧阳缮, 廖可非, 等. 基于深度学习的GPR B-SCAN图像双曲线检测方法[J]. 电子学报, 2021, 49(5): 953–963. doi: 10.12263/DZXB.20200635 WANG Hui, OUYANG Shan, LIAO Kefei, et al. GPR B-SCAN image hyperbola detection method based on deep learning[J]. Acta Electronica Sinica, 2021, 49(5): 953–963. doi: 10.12263/DZXB.20200635
[16]	REN Shaoqing, HE Kaiming, 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
[17]	杨必胜, 宗泽亮, 陈驰, 等. 车载探地雷达地下目标实时探测法[J]. 测绘学报, 2020, 49(7): 874–882. doi: 10.11947/j.AGCS.2020.20190293 YANG Bisheng, ZONG Zeliang, CHEN Chi, et al. Real time approach for underground objects detection from vehicle-borne ground penetrating radar[J]. Acta Geodaetica et Cartographica Sinica, 2020, 49(7): 874–882. doi: 10.11947/j.AGCS.2020.20190293
[18]	REDMON J and FARHADI A. YOLOv3: An incremental improvement[EB/OL].https://arxiv.org/abs/1804.02767, 2018.
[19]	ZHANG Jun, YANG Xing, LI Weiguang, et al. Automatic detection of moisture damages in asphalt pavements from GPR data with deep CNN and IRS method[J]. Automation in Construction, 2020, 113: 103119. doi: 10.1016/j.autcon.2020.103119
[20]	HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al. Deep residual learning for image recognition[C]. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, USA, 2016: 770–778.
[21]	REDMON J and FARHADI A. YOLO9000: Better, faster, stronger[C]. 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, USA, 2017: 6517–6525.
[22]	BRUNZELL H. Detection of shallowly buried objects using impulse radar[J]. IEEE Transactions on Geoscience and Remote Sensing, 1999, 37(2): 875–886. doi: 10.1109/36.752207
[23]	WARREN C, GIANNOPOULOS A, and GIANNAKIS I. gprMax: Open source software to simulate electromagnetic wave propagation for ground penetrating radar[J]. Computer Physics Communications, 2016, 209: 163–170. doi: 10.1016/j.cpc.2016.08.020
[24]	DÉROBERT X and PAJEWSKI L. TU1208 open database of radargrams: The dataset of the IFSTTAR geophysical test site[J]. Remote Sensing, 2018, 10(4): 530. doi: 10.3390/rs10040530