高级搜索

面向GPR B-scan图像的目标检测算法综述

侯斐斐 施荣华 雷文太 董健 许孟迪 席景春

引用本文: 侯斐斐, 施荣华, 雷文太, 董健, 许孟迪, 席景春. 面向GPR B-scan图像的目标检测算法综述[J]. 电子与信息学报, doi: 10.11999/JEIT190680 shu
Citation:  Feifei HOU, Ronghua SHI, Wentai LEI, Jian DONG, Mengdi XU, Jingchun XI. A Review of Target Detection Algorithm for GPR B-scan Processing[J]. Journal of Electronics and Information Technology, doi: 10.11999/JEIT190680 shu

面向GPR B-scan图像的目标检测算法综述

    作者简介: 侯斐斐: 女,1993年生,博士生,研究方向为探地雷达,深度学习,图像处理;
    施荣华: 男,1963年生,教授,博士生导师,研究方向为射频系统集成和量子技术;
    雷文太: 男,1979年生,副教授,博士生导师,研究方向为探地雷达系统集成和信号处理;
    通讯作者: 雷文太,leiwentai@csu.edu.cn
  • 基金项目: 国家自然科学基金(61102139和61872390),中南大学基础研究基金(2018zzts181)

摘要: 利用无损探测技术来获取地下目标的信息是当前研究的热点,探地雷达(GPR)作为一种重要的无损工具,已被广泛用于检测,定位和特征化地下目标。然而,从GPR成像中探测掩埋物体并评估其位置既费时又费力。因此,实现地下目标的自动化探测对实际应用是必要的。为此,该文在综合分析地下目标回波特征的基础上,讨论了使用GPR评估目标位置的可行性,并回顾了国内外学者在GPR成像中对双曲线特征自动化检测的研究进展。该文还在国内外典型实例剖析的基础上,总结并比较了目标检测的处理方法。最后指出,未来的研究应集中于开发新的深度学习检测框架,用以自动检测和估计真实场景中的地下特征。

English

    1. [1]

      JOL H M, 雷文太, 童孝忠, 周旸, 译. 探地雷达理论与应用[M]. 北京: 电子工业出版社, 2011.
      JOL H M, LEI Wentai, TONG Xiaozhong, ZHOU Yang, translation. Ground Penetrating Radar: Theory and Applications[M]. Beijing: Publishing House of Electronics Industry, 2011.

    2. [2]

      BESAW L E and STIMAC P J. Deep convolutional neural networks for classifying GPR B-Scans[C]. SPIE 9454, Detection and Sensing of Mines, Explosive Objects, and Obscured Targets XX, Baltimore, USA, 2015: 945413.

    3. [3]

      BESAW L E. Detecting buried explosive hazards with handheld GPR and deep learning[C]. SPIE 9823, Detection and Sensing of Mines, Explosive Objects, and Obscured Targets XXI, Baltimore, USA, 2016: 98230N. doi: 10.1117/12.2223797.

    4. [4]

      BRALICH J, REICHMAN D, COLLINS L M, et al. Improving convolutional neural networks for buried target detection in ground penetrating radar using transfer learning via pretraining[C]. SPIE 10182, Detection and Sensing of Mines, Explosive Objects, and Obscured Targets XXⅡ, Anaheim, USA, 2017.

    5. [5]

      REICHMAN D, COLLINS L M, and MALOF J M. Some good practices for applying convolutional neural networks to buried threat detection in Ground Penetrating Radar[C]. The 2017 9th International Workshop on Advanced Ground Penetrating Radar, Edinburgh, UK, 2017: 1–5.

    6. [6]

      LAMERI S, LOMBARDI F, BESTAGINI P, et al. Landmine detection from GPR data using convolutional neural networks[C]. The 2017 25th European Signal Processing Conference, Kos, Greece, 2017: 508–512.

    7. [7]

      BENEDETTO A, BENEDETTO F, DE BLASⅡS M R, et al. Reliability of radar inspection for detection of pavement damage[J]. Road Materials and Pavement Design, 2004, 5(1): 93–110. doi: 10.1080/14680629.2004.9689964

    8. [8]

      LEI Wentai, SHI Ronghua, DONG Jian, et al. A multi-scale weighted back projection imaging technique for ground penetrating radar applications[J]. Remote Sensing, 2014, 6(6): 5151–5163. doi: 10.3390/rs6065151

    9. [9]

      LEI Wentai, ZENG Sheng, ZHAO Jian, et al. An improved back projection imaging algorithm for subsurface target detection[J]. Turkish Journal of Electrical Engineering & Computer Sciences, 2013, 21(6): 1820–1826.

    10. [10]

      BENEDETTO F and TOSTI F. GPR spectral analysis for clay content evaluation by the frequency shift method[J]. Journal of Applied Geophysics, 2013, 97: 89–96. doi: 10.1016/j.jappgeo.2013.03.012

    11. [11]

      KAUR P, DANA K J, ROMERO F A, et al. Automated GPR rebar analysis for robotic bridge deck evaluation[J]. IEEE Transactions on Cybernetics, 2016, 46(10): 2265–2276. doi: 10.1109/TCYB.2015.2474747

    12. [12]

      DINH K, GUCUNSKI N, and DUONG T H. An algorithm for automatic localization and detection of rebars from GPR data of concrete bridge decks[J]. Automation in Construction, 2018, 89: 292–298. doi: 10.1016/j.autcon.2018.02.017

    13. [13]

      YUAN Chenxi, LI Shuai, CAI Hubo, et al. GPR signature detection and decomposition for mapping buried utilities with complex spatial configuration[J]. Journal of Computing in Civil Engineering, 2018, 32(4): 04018026. doi: 10.1061/(ASCE)CP.1943-5487.0000764

    14. [14]

      LI Shuai, CAI Hubo, and KAMAT V R. Uncertainty-aware geospatial system for mapping and visualizing underground utilities[J]. Automation in Construction, 2015, 53: 105–119. doi: 10.1016/j.autcon.2015.03.011

    15. [15]

      LI Shuai, CAI Hubo, ABRAHAM D M, et al. Estimating features of underground utilities: Hybrid GPR/GPS approach[J]. Journal of Computing in Civil Engineering, 2016, 30(1): 04014108. doi: 10.1061/(ASCE)CP.1943-5487.0000443

    16. [16]

      ILLINGWORTH J and KITTLER J. A survey of the Hough transform[J]. Computer Vision, Graphics, and Image Processing, 1988, 43(2): 280.

    17. [17]

      BORGIOLI G, CAPINERI L, FALORNI P, et al. The detection of buried pipes from time-of-flight radar data[J]. IEEE Transactions on Geoscience and Remote Sensing, 2008, 46(8): 2254–2266. doi: 10.1109/tgrs.2008.917211

    18. [18]

      WINDSOR C G, CAPINERI L, and FALORNI P. A data pair-labeled generalized Hough transform for radar location of buried objects[J]. IEEE Geoscience and Remote Sensing Letters, 2014, 11(1): 124–127. doi: 10.1109/LGRS.2013.2248119

    19. [19]

      BOOKSTEIN F L. Fitting conic sections to scattered data[J]. Computer Graphics and Image Processing, 1979, 9(1): 56–71. doi: 10.1016/0146-664x(79)90082-0

    20. [20]

      AKIMA H. A method of bivariate interpolation and smooth surface fitting for irregularly distributed data points[J]. ACM Transactions on Mathematical Software, 1978, 4(2): 148–159. doi: 10.1145/355780.355786

    21. [21]

      PORRILL J. Fitting ellipses and predicting confidence envelopes using a bias corrected Kalman filter[J]. Image and Vision Computing, 1990, 8(1): 37–41. doi: 10.1016/0262-8856(90)90054-9

    22. [22]

      YOUN H S and CHEN C C. Automatic GPR target detection and clutter reduction using neural network[C]. SPIE 4758, 9th International conference on Ground Penetrating Radar, Santa Barbara, USA, 2002: 579–582. doi: 10.1117/12.462229.

    23. [23]

      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

    24. [24]

      VIOLA P and JONES M J. Robust real-time face detection[J]. International Journal of Computer Vision, 2004, 57(2): 137–154. doi: 10.1023/b:visi.0000013087.49260.fb

    25. [25]

      KRIZHEVSKY A, SUTSKEVER I, and HINTON G E. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60(6): 84–90. doi: 10.1145/3065386

    26. [26]

      WITTEN T R. Present state of the art in ground-penetrating radars for mine detection[C]. SPIE 3392, Detection and Remediation Technologies for Mines and Minelike Targets Ⅲ, Orlando, USA, 1998.

    27. [27]

      PHAM M T and LEFÈVRE S. Buried object detection from B-scan ground penetrating radar data using Faster-RCNN[C]. Proceedings of 2018 IEEE International Geoscience and Remote Sensing Symposium, Valencia, Spain, 2018: 6804–6807.

    28. [28]

      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

    29. [29]

      DOU Qingxu, WEI Lijun, MAGEE D R, et al. Real-time hyperbola recognition and fitting in GPR data[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(1): 51–62. doi: 10.1109/tgrs.2016.2592679

    30. [30]

      ZHOU Xiren, CHEN Huanhuan, and LI Jinlong. An automatic GPR B-Scan image interpreting model[J]. IEEE Transactions on Geoscience and Remote Sensing, 2018, 56(6): 3398–3412. doi: 10.1109/TGRS.2018.2799586

    31. [31]

      TANOLI, W A, SHARAFAT, A, and Park J, et al. Damage Prevention for underground utilities using machine guidance[J]. Automation in Construction, 2017, 107: 102893.

    32. [32]

      YALÇINER C C, BANO M, KADIOGLU M, et al. New temple discovery at the archaeological site of Nysa (western Turkey) using GPR method[J]. Journal of Archaeological Science, 2009, 36(8): 1680–1689. doi: 10.1016/j.jas.2008.12.016

    33. [33]

      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

    34. [34]

      PASOLLI E, MELGANI F, DONELLI M, et al. Automatic detection and classification of buried objects in GPR images using genetic algorithms and support vector machines[C]. 2008 IEEE International Geoscience and Remote Sensing Symposium, Boston, USA, 2008: Ⅱ-525–Ⅱ-528.

    35. [35]

      PASOLLI E, MELGANI F, and DONELLI M. Automatic analysis of GPR images: A pattern-recognition approach[J]. IEEE Transactions on Geoscience and Remote Sensing, 2009, 47(7): 2206–2217. doi: 10.1109/TGRS.2009.2012701

    36. [36]

      MOLYNEAUX T C K, MILLARD S G, BUNGEY J H, et al. Radar assessment of structural concrete using neural networks[J]. NDT & E International, 1995, 28(5): 281–288.

    37. [37]

      AL-NUAIMY W, HUANG Y, NAKHKASH M, et al. Automatic detection of buried utilities and solid objects with GPR using neural networks and pattern recognition[J]. Journal of Applied Geophysics, 2000, 43(2/4): 157–165.

    38. [38]

      GAMBA P and LOSSANI S. Neural detection of pipe signatures in ground penetrating radar images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2000, 38(2): 790–797. doi: 10.1109/36.842008

    39. [39]

      AL-NUAIMY W, HUANG Y, NAKHKASH M, et al. Neural network for the automatic detection of buried utilities and landmines[C]. 1998 Progress of Electromagnetic Research Symposium, Nantes, Frances, 1998: 141.

    40. [40]

      SHIHAB S, AL-NUAIMY W, HUANG Y, et al. Automatic region-based shape discrimination of ground penetrating radar signatures[C]. 2003 Symposium on the Application of Geophysics to Environmental and Engineering Problems SAGEEP 2003, San Antonio, USA, 2003.

    41. [41]

      AL-NUAIMY W, LU Huihai, SHIHAB S, et al. Automatic mapping of linear structures in 3-dimensional space from ground-penetrating radar data[C]. 2001 IEEE/ISPRS Joint Workshop on Remote Sensing and Data Fusion over Urban Areas, Rome, Italy, 2001: 198–201.

    42. [42]

      SHIHAB S, AL-NUAIMY W, and ERIKSEN A. Image processing and neural network techniques for automatic detection and interpretation of ground penetrating radar data[C]. The 6th WSEAS, Crete, 2002.

    43. [43]

      AL-NUAIMY W, HUANG Yi, ERIKSEN A, et al. Automatic detection of hyperbolic signatures in ground-penetrating radar data[C]. SPIE 4491, Subsurface and Surface Sensing Technologies and Applications Ⅲ, San Diego, USA, 2001: 327.

    44. [44]

      SHAW M R, MOLYNEAUX T C K, MILLARD S G, et al. Assessing bar size of steel reinforcement in concrete using ground penetrating radar and neural networks[J]. Insight - Non-Destructive Testing and Condition Monitoring, 2003, 45(12): 813–816. doi: 10.1784/insi.45.12.813.52980

    45. [45]

      SHAW M R, MILLARD S G, MOLYNEAUX T C K, et al. Location of steel reinforcement in concrete using ground penetrating radar and neural networks[J]. NDT & E International, 2005, 38(3): 203–212.

    46. [46]

      LAMERI S, LOMBARDI F, BESTAGINI P, et al. Landmine detection from GPR data using convolutional neural networks[C]. The 2017 25th European Signal Processing Conference, Kos, Greece, 2017.

    47. [47]

      BENGIO Y. Learning deep architectures for AI[J]. Foundations and Trends® in Machine Learning, 2009, 2(1): 1–127. doi: 10.1561/2200000006

    48. [48]

      REICHMAN D, COLLINS L M, and MALOF J M. Some good practices for applying convolutional neural networks to buried threat detection in Ground Penetrating Radar[C]. The 2017 9th International Workshop on Advanced Ground Penetrating Radar, Edinburgh, UK, 2017.

    49. [49]

      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

    50. [50]

      KRIZHEVSKY A. Learning multiple layers of features from tiny images[R]. Technical Report TR-2009, 2009: 1–60.

    51. [51]

      KANUNGO T, MOUNT D M, NETANYAHU N S, et al. An efficient k-means clustering algorithm: Analysis and implementation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(7): 881–892. doi: 10.1109/TPAMI.2002.1017616

    52. [52]

      NG R T and HAN Jiawei. Efficient and effective clustering methods for spatial data mining[C]. The 20th International Conference on Very Large Data Bases, Birmingham, USA, 1994: 144–155.

    53. [53]

      ESTER M, KRIEGEL H P, SANDER J, et al. A density-based algorithm for discovering clusters a density-based algorithm for discovering clusters in large spatial databases with noise[C]. The 2nd International Conference on Knowledge Discovery and Data Mining, Portland, Oregon, 1996: 226–231.

    54. [54]

      ERTÖZ L, STEINBACH M, and KUMAR V. Finding clusters of different sizes, shapes, and densities in noisy, high dimensional data[C]. The 2nd SIAM International Conference on Data Mining, 2003.

    55. [55]

      AHN S J, RAUH W, and WARNECKE H J. Least-squares orthogonal distances fitting of circle, sphere, ellipse, hyperbola, and parabola[J]. Pattern Recognition, 2001, 34(12): 2283–2303. doi: 10.1016/S0031-3203(00)00152-7

    56. [56]

      FITZGIBBON A, PILU M, and FISHER R B. Direct least square fitting of ellipses[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1999, 21(5): 476–480. doi: 10.1109/34.765658

    57. [57]

      GANDER W, GOLUB G H, and STREBEL R. Least-squares fitting of circles and ellipses[J]. BIT Numerical Mathematics, 1994, 34(4): 558–578. doi: 10.1007/BF01934268

    58. [58]

      PILU M, FITZGIBBON A W, and FISHER R B. Ellipse-specific direct least-square fitting[C]. The 3rd IEEE International Conference on Image Processing, Lausanne, Switzerland, 1996: 599–602.

    59. [59]

      CHEN Huanhuan and COHN A G. Probabilistic conic mixture model and its applications to mining spatial ground penetrating radar data[C]. The Workshops in SIAM Conference on Data Mining, 2010: 1–9.

    60. [60]

      CHEN Huanhuan and COHN A G. Probabilistic robust hyperbola mixture model for interpreting ground penetrating radar data[C]. 2010 International Joint Conference on Neural Networks, Barcelona, Spain, 2010: 1–8.

    1. [1]

      董书琴, 张斌. 基于深度特征学习的网络流量异常检测方法. 电子与信息学报,

    2. [2]

      冯浩, 黄坤, 李晶, 高榕, 刘东华, 宋成芳. 基于深度学习的混合兴趣点推荐算法. 电子与信息学报,

    3. [3]

      杨宏宇, 王峰岩. 基于深度卷积神经网络的气象雷达噪声图像语义分割方法. 电子与信息学报,

    4. [4]

      张烨, 许艇, 冯定忠, 蒋美仙, 吴光华. 基于难分样本挖掘的快速区域卷积神经网络目标检测研究. 电子与信息学报,

    5. [5]

      唐伦, 赵培培, 赵国繁, 陈前斌. 基于深度信念网络资源需求预测的虚拟网络功能动态迁移算法. 电子与信息学报,

    6. [6]

      袁野, 贾克斌, 刘鹏宇. 基于深度卷积神经网络的多元医学信号多级上下文自编码器. 电子与信息学报,

    7. [7]

      钱亚冠, 卢红波, 纪守领, 周武杰, 吴淑慧, 云本胜, 陶祥兴, 雷景生. 基于粒子群优化的对抗样本生成算法. 电子与信息学报,

    8. [8]

      徐金甫, 吴缙, 李军伟, 曲彤洲, 董永兴. 基于敏感度混淆机制的控制型物理不可克隆函数研究. 电子与信息学报,

    9. [9]

      谢湘, 张立强, 王晶. 残差网络在婴幼儿哭声识别中的应用. 电子与信息学报,

    10. [10]

      桑海峰, 陈紫珍. 基于双向门控循环单元的3D人体运动预测. 电子与信息学报,

    11. [11]

      陈鸿昶, 吴彦丞, 李邵梅, 高超. 基于行人属性分级识别的行人再识别. 电子与信息学报,

    12. [12]

      刘政怡, 徐天泽. 基于优化的极限学习机和深度层次的RGB-D显著检测. 电子与信息学报,

    13. [13]

      王斐, 吴仕超, 刘少林, 张亚徽, 魏颖. 基于脑电信号深度迁移学习的驾驶疲劳检测. 电子与信息学报,

    14. [14]

      兰巨龙, 于倡和, 胡宇翔, 李子勇. 基于深度增强学习的软件定义网络路由优化机制. 电子与信息学报,

    15. [15]

      王鑫, 李可, 宁晨, 黄凤辰. 基于深度卷积神经网络和多核学习的遥感图像分类方法. 电子与信息学报,

    16. [16]

      李瑞, 张群, 苏令华, 梁佳, 罗迎. 基于稀疏贝叶斯学习的双基雷达关联成像. 电子与信息学报,

    17. [17]

      郭晨, 简涛, 徐从安, 何友, 孙顺. 基于深度多尺度一维卷积神经网络的雷达舰船目标识别. 电子与信息学报,

    18. [18]

      徐少毅, 高帅. 机器对机器通信中一种基于能量效率与系统容量的多目标无线资源管理算法. 电子与信息学报,

    19. [19]

      魏嘉琪, 张磊, 刘宏伟, 盛佳恋. 曲线交叠外推的微动多目标宽带分辨算法. 电子与信息学报,

    20. [20]

      胡长雨, 汪玲, 朱栋强. 结合字典学习技术的ISAR稀疏成像方法. 电子与信息学报,

  • 图 1  真实场景中GPR B-scan中的双曲线特征(来源于美国田纳西大学Newland Greenway桥面)

    图 2  在文献[28]中解决的一些复杂情况示例

    图 3  DCSE和OSCA算法的对比结果展示

    图 4  基于CTFP算法的双曲线拟合结果展示

    表 1  GPR目标检测的经典算法总结

    序号参考文献时间GPR目标客观评价
    1Borgioli et al. [17]2008地埋管道在Hough变换中引入加权因子,解决了管道靠近时双曲线重叠的问题;但是需要预备模型,计算成本相对较高。
    2Maas et al. [23]2013双曲线反射使用Viola-Jones算法标记目标候选区域,它避免了模板匹配并缩小了后续搜索区域;然而,应用特征需手动识别,分类结果取决于特征的质量,难度随着数据量的增加。
    3Besaw et al. [2]2016地埋爆炸物应用CNN从GPR B-scan中提取有意义的特征并对目标进行分类。交叉验证,网络权重正则化和“dropout”用于防止过度训练。
    4Bishop et al. [3]2016地埋爆炸物在CNN基础上增加了额外的Data Augmentation技术,用于增加可用训练数据的数量和可变性。
    5Daniel et al. [4,5]2017地埋爆炸物研究了预训练CNN的初始化步骤,以解决GPR数据标记样本不足的问题;但是输入网络中真实图像的大小和数量通常是有限的,仅实现分类步骤。
    6Pham et al. [27]2018双曲线反射首次采用Faster RCNN来检测GPR B-scan中的反射双曲线。该技术在真实测试集上的性能要超过使用HOG或Haar-like特征的检测器,但缺少定量的评估。
    7Hou et al. [28]2019地埋钢筋在文献[27]基础上,采用了DA手段增加真实GPR数据集和仿真数据集;提出DCSE算法以识别双曲线特征,完善了文献[30]中提出的OSCA算法;提出CTFP算法自动提取拟合点。所提出方案的有效性在仿真和真实数据集上得到了验证。
    8Doue et al. [29]2016双曲线反射提出了C3算法分割交叉双曲线,并将其送入神经网络进行分类。C3算法水平扫描B-scan图像中的每个像素以进行聚类。然而,双曲线是垂直向下打开的,C3算法没有考虑这个重要特征。
    9Zhou et al. [30]2018金属管道
    水泥管道
    提出OSCA算法解决了文献[29]中的难题,可以识别具有向下开口特征的聚类。然而,在整个图像上进行OSCA算法是不合适的,因为难以处理包含太多非平稳噪声的大型现场数据集,导致后续处理复杂化。
    下载: 导出CSV
  • 加载中
图(4)表(1)
计量
  • PDF下载量:  24
  • 文章访问数:  170
  • HTML全文浏览量:  137
文章相关
  • 通讯作者:  雷文太, leiwentai@csu.edu.cn
  • 录用日期:  2019-11-12
  • 网络出版日期:  2019-11-18
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

/

返回文章