Designing an Applied Algorithm to Extract Geometrical Parameters of Buried Cylindrical Targets in Ground-Penetrating Radar (GPR) Method

Document Type : Research Article

Authors

1 Dept. of Mining, Tehran University, Iran

2 Dept. of Mining, Arak University of Technology, Iran

10.29252/anm.2019.8418.1290

Abstract

Summary
In the present research, GPR method has been used to identify geometrical parameters of buried cylindrical targets. To achieve this goal, first, forward modeling of GPR data has been carried out for several synthetic models corresponding to common targets in geotechnical and subsurface installations. Then, an applied algorithm on the basis of signal processing on a radargram (A Scan) with high accuracy was implemented in MATLAB environment. The performance of the algorithm was validated for several synthetic models so that led to the favorite results in all cases. The primitive algorithm was improved to employ for real GPR images having a large amount of various noises.
 
Introduction
One of the most important tasks in engineering design is extraction of geometrical parameters of subsurface hidden objects. In this research, has been attempted to investigate the treatment of GPR responses in spatial domain using simulated response of various synthetic models by forward modeling. Then after extraction of relationships between existence mathematical models with GPR system response, geometrical parameters of buried cylindrical targets and physical parameters of host medium are identified through employing appropriate algorithms and image processing methods.
 
Methodology and Approaches
An applied algorithm based on signal processing on a radargram (A Scan) was implemented in MATLAB environment which investigates the treatment of GPR response in spatial domain. The performance of the algorithm was validated for several synthetic models such as empty metallic and PVC horizontal cylinders and also the model including the couple of empty horizontal cylinders made by PVC. The algorithm was improved by applying Cascade Object Detector (COD) algorithm, interest of region is defined so that null regions are removed, implementing on the finite interested region. Then the algorithm is trained based on definition of positive and negative images. The performance of the proposed algorithm was evaluated for a real GPR radargram related to one of the profiles surveyed in Imam-Hossein square, opposite the municipality of Isfahan city so that also yielded a favorite result in this regard.
 
Results and Conclusions
In order to evaluate the accuracy of curve fitting the hyperbolic equation on the data, statistical validation criterion well-known as determination coefficient has been used. According to this criterion fitting accuracy of hyperbolic equation on the data for all synthetic models except the couple of horizontal cylinders is up to 93 percent. The algorithm has estimated the geometrical parameters of cylindrical targets with an error less than 8 percent. Also using the improved algorithm, determination coefficient of the fitted curve is 83.99 percent that is a favorite result. The algorithm could estimate radius, burial depth and horizontal location of the buried horizontal cylinder in the real GPR image with the errors of 7.6, 1.7 and 1.1 percent, respectively.

Keywords

Main Subjects

Full Text

تعیین پارامترهای هندسی و فیزیکی اشیاء مدفون زیرسطحی در اغلب مطالعات از جمله ژئوتکنیک و شناسایی ساختار، باستان‌شناسی، تاسیساتی و اکتشافی به عنوان یک هدف مطرح است. در واقع یکی از مهم‌ترین مسائل در طراحی‌ها و تصمیم­گیری‌های مهندسی، شناسایی و استخراج پارامترهای هندسی ناهمگنی‌های زیرسطحی مدفون است که تاحدودی با استفاده از روش‌های ژئوفیزیکی مانند GPR قابل حل است. بنابراین در این پژوهش سعی شده است تا با استفاده از پاسخ شبیه‌سازی شده مدل‌های مصنوعی مختلف به روش مدلسازی پیشرو، رفتار پاسخ‌های حاصل در حوزه مکانی[i] مورد بررسی و مطالعه قرار گیرد. در مرحله بعد پس از استخراج رابطه‌های پنهان بین مدل‌های ریاضی موجود و پاسخ سیستم GPR حاصل، از طریق به کارگیری الگوریتم‌های مناسب و روش‌های پردازش تصویر، اقدام به شناسایی پارامترهای هندسی اهداف استوانه‌ای و فیزیکی محیط میزبان مدفون می‌شود. از آنجایی که اشیاء استوانه‌ای در پاسخ GPR توسط پارامترهایی همانند شعاع، عمق دفن، موقعیت افقی و نیز ویژگی‌های محیط میزبان همانند سرعت سیر امواج الکترومغناطیسی مشخص می‌شوند، بنابراین با استفاده از مدل ریاضی هذلولی می‌توان رابطه‌های بین پارامترهای هذلولی پاسخ و پارامترهای هندسی اشیاء مدفون و فیزیکی محیط میزبان را استخراج نمود.

در این پژوهش ابتدا برای استخراج پارامترهای هندسی اهداف استوانه‌ای زیرسطحی و سرعت سیر موج الکترومغناطیس در محیط در برگیرنده مدل‌های مصنوعی، یک الگوریتم کاربردی در محیط نرم‌افزار MATLAB پیاده‌سازی شد. در ادامه الگوریتم کامل‌تری معرفی می‌شود که اساس ساختار آن را الگوریتم پیشنهادی قبلی تشکیل داده و برای تصاویر واقعی، بهبود یافته است. خطای تخمین پارامترهای هندسی بر روی مدل‌های مصنوعی و واقعی در همه موارد مقدار قابل قبولی بوده و الگوریتم در بیش‌تر موارد با خطای زیر ده درصد پارامترهای هندسی اهداف استوانه‌ای را تخمین زده است.


1- Spatialdomain

References

]1[          Ahmadi, R., Fathianpour, N., & Norouzi, G.H. (2016). Detecting cylindrical targets characteristics hidden in GPR images using artificial neural network and template matching, Journal of Engineering Geology, 9(4), 3069-3092 (In Persian).
]2[          Ahmadi, R., Fathianpour, N., & Norouzi, G.H. (2016). Comparison of the performance of ANN and SVM methods in automatic detection of hidden cylindrical targets in GPR images, Iranian Journal of Mining Engineering (IRJME), 10(26), 83-98 (In Persian).
]3[          Hosseini, M., Kamkar-Rouhani, A., Mohammadi-Vizhe, M., & Parnow, S. (2017). Importance of three-dimensional ground penetrating radar surveys in accurate display of the images of subsurface installations, Case study: Shahrood University of Technology, Journal of Research on Applied Geophusics (JRAG), 2(2), 67-76 (In Persian).
[4]        Shihab, S., & Al-Nuaimy, W. (2005). Radius estimation for cylindrical objects detected by ground- penetrating radar, Subsurface Sensing Technologies and Applications, Vol. 6, pp. 151-166.
[5]           Toksoz, D., Yilmaz, I., Seren, A., & Mataraci, I. (2016). A study on the performance of GPR for detection of different types of buried objects, Procedia Engineering, Vol. 161, pp. 399–406.
[6]           Zhang, P., Guo, X., Muhammat, N., & Wang, X. (2016). Research on probing and predicting the diameter of an underground pipeline by GPR during an operation period, Tunnelling and Underground Space Technology, Vol. 58, pp. 99–10
[7]           Ahmadi, R., & Fathianpour, N. (2017). Estimating geometrical parameters of cylindrical targets detected by Ground-Penetrating Radar using template matching algorithm, Arab J Geosci, pp. 10: 140.
[8]        Sadiku, M.N.O. (2001). Numerical techniques in electromagnetics. Second edition, Boca Raton London New York Washington, D.C. CRC press.
[9]        Annan, A.P. (2002). The history of ground-penetrating radar, Subsurface Sensing Technologies and Applications, 3(4), pp. 303–320.
[10]         Annan, A.P. (2001). Ground-penetrating radar workshop notes, Sensors and Software Inc, Mississauga, ON, Canada, 192 pages.
[11]       Annan, A.P. (2003). Ground- penetrating radar: Principles, procedures & applications, Sensors & Software Inc. Technical Paper.
[12]       Annan, A.P., & Cosway, S.W. (1992). Ground-penetrating radar survey design, Proceedings of the Symposium on the Application of Geophysics to Engineering and Environmental Problems, SAGEEP’92, April 26-29, Oakbrook, IL, pp. 329-351.
[13]       Poluha, B., Porsani, J.L., Almeida, E.R., Santos, V., & Allen, S.J. (2017). Depth estimates of buried utility systems using the GPR method: studies at the IAG/USP geophysics test site, DOI: 10.4236/ijg.2017.85040.
[14]         Davis, J.L., & Annan, A.P. (1989). Ground-penetrating radar for high-resolution mapping of soil and rock stratigraphy, Geophysical Prospecting, 37, pp. 531-551
[15]       Viola, P., & Michael, J. J. (2001). Rapid object detection using a boosted cascade of simple features, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Volume: 1, pp. 511–518.
[16]       Lienhart, R., Kuranov, A., & Pisarevsky, V. (2003). Empirical analysis of detection cascades of boosted classifiers for rapid object detection, Proceedings of the 25th DAGM Symposium on Pattern Recognition. Magdeburg, Germany.
[17]       Ojala, T., Pietikäinen. M., & Mäenpää, T. (2002). Multiresolution gray-scale and rotation invariant texture classification with local binary patterns, In IEEE Transactions on Pattern Analysis and Machine Intelligence. Volume 24, Issue 7, pp. 971-987.
[18]         Kruppa, H., Castrillon-Santana, M., & Schiele, B. (2003). Fast and robust face finding via local context, Proceedings of the Joint IEEE International Workshop on Visual Surveillance and Performance Evaluation of Tracking and Surveillance, pp. 157–164.
[19]         Dalal, N., & Triggs, B. (2005). Histograms of oriented gradients for human detection, IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Volume 1, pp. 886–893.
Journal of Analytical and Numerical Methods in Mining Engineering
Volume 9, Issue 19
September 2019
Pages 63-78

Files

History

  • Receive Date: 17 October 2017
  • Revise Date: 24 December 2018
  • Accept Date: 25 December 2018

Share

How to cite

Statistics