CC BY
27TECHNICAL SCIENCES
ПРИМЕНЕНИЕ КОРРЕЛЯЦИОННОГО АНАЛИЗА ДЛЯ ОБРАБОТКИ СИГНАЛОВ ГЕОРАДАРА В СПЕКТРАЛЬНОЙ ОБЛАСТИ
Абрамович А.О.
Национальный технический университет Украины «Киевский политехнический институт имени Игоря Сикорского», г. Киев, Украина, соискатель
Поддубный В. О.
Национальный технический университет Украины «Киевский политехнический институт имени Игоря Сикорского», г. Киев, Украина к.т.н., доцент
Баженов В.Г.
Национальный технический университет Украины «Киевский политехнический институт имени Игоря Сикорского», г. Киев, Украина к.т.н., доцент
APPLICATION OF CORRELATION ANALYSIS FOR PROCESSING GEORADAR'S SIGNALS IN
THE SPECTRAL REGION
Abramovych A.
National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" Kyiv, Ukraine, applicant
Piddubnyi V.
National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" Kyiv, Ukraine, assistant professor Bazhenov V.
National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" Kyiv, Ukraine, assistant professor
АННОТАЦИЯ
Предложен метод повышения достоверности распознавания металлических объектов, обнаруженных георадаром. Исследования проводились на примере георадара, работающего по вихретоковому принципу действия и имеющего низкочастотные магнитные антенны. Предложена корреляционная обработка спектра сигнала отклика от скрытого объекта, что позволяет повысить надежность идентификации металла при обработке сигнала в спектральной области на основе преобразования Фурье. Исследования проводились на примере металлов с близкими спектральными характеристиками.
ABSTRACT
A method has been proposed to increase the reliability of material recognition of metal objects detected by georadar. The research was carried out on the example of a georadar operating on the eddy current principle of operation and having low-frequency magnetic antennas. Correlation processing of the response signal spectrum from a hidden object is proposed, which allows to increase the reliability of metal identification when processing the signal in the spectral region based on the Fourier transform. Studies have been carried out on the example of metals with similar spectral characteristics.
Ключевые слова: Вихретоковая система и устройства, анализ металлов, корреляция, идентификация металлов.
Keywords: Eddy current system and device, metal analysis, correlation, identification of metals.
Electronic devices that detect and identify metal objects are used to locate and identify hidden metal objects in different environments that differ in their physical characteristics. They use the differences in the electrical and magnetic properties of the objects themselves and the environment in which they are located.
Georadars (GPR) carry out non-destructive monitoring of road conditions, are used in geological and archaeological explorations, as well as for the detection of metal objects. The simplest in design is a georadar with a magnetic frame antenna, which is essentially an eddy current device. Distinguishing objects in them is carried out by the threshold level of the received signal. However, this approach does not allow you to determine the material from which the hidden object is
made. To overcome the described limitations, a method of distinction is proposed, which is based on the use of spectral analysis of the signal that occurs in the receiving antenna of GPR under the action of Foucault currents.
The task of identifying hidden metal objects consists of two stages: the detection of a metal object and the identification of the metal from which the object is made.
Detection is carried out by radio systems - geora-dars [1-3], which work on the basis of irradiation of the hidden object with an electromagnetic field and subsequent analysis of the signal obtained as a result of the
interaction of the field with the object. The disadvantage of such devices is the inability to determine the metal from which the object is made.
The technical characteristics of eddy current devices are determined by the structure of the device and the methods of processing the signal received from the object. The information about the metal lies in the amplitude, phase or frequency of the signal, and, as a rule, only one of these parameters is informative and the others are considered interfering. In existing devices, the information parameter is usually the signal amplitude [1].
Phase and frequency methods are usually used in non-destructive testing devices for metal objects to detect mechanical defects in the object. Although in a number of works [4,5] an attempt was made to use the phase method to detect and identify the metal, but the authors were able to obtain identification only by di-chotomous features (ferrous /non ferrous).
The type of detected metal can be determined only visually, but for this purpose the object must be analyzed in the laboratory on special metal analyzers.
There are several methods of such analysis [6]. The most common chemical method, which is based on the decomposition of the studied material into atoms and molecules and the subsequent study of the spectrum obtained from the constituent parts of the object. This is the most accurate method of analysis. But it is laboratory and time consuming.
Now the chemical method is replaced by optical emission, X-ray fluorescence, atomic absorption, atomic emission and other methods. The most common of these are optical emission and X-ray fluorescence analyzes. These methods require the presence of a metal sample [7] and also do not allow to determine its composition remotely, without laboratory tests.
Therefore, the task of detecting and analyzing the composition of metal objects without destroying them, including those hidden in another dielectric medium, is relevant.
Modern georadars are based on the analysis of the interaction of the electromagnetic field of the antenna with a metal object of control on the basis of Foucault eddy currents induced by the electromagnetic field. There are now a large number of radars [2], which are based on the eddy current method [8,9].
The change in the electromagnetic properties of a medium or a hidden object, in which the propagation of an exciting electromagnetic field occurs, leads to a
change in the parameters of the vortex currents and the generated secondary magnetic flux. As a result, the total magnetic flux of the system "eddy current converter - a hidden object" changes, which, in turn, leads to a change in the electrical parameters (resistance, current, ether) in the receiving antenna. These changes are monitored in the instruments included in the measuring range. An eddy current converter (transmitting antenna) is generally a coil, creating a homogeneous variable magnetic field. In the search for a hidden object, the induction converter moves above the object's surface. When crossing the field and the metal object, the magnetic flux is distorted and an electric signal appears in the sensitive element of the device (receiving antenna), which carries information about the presence of the object in a given location. As a rule, the antenna system consists of transmitting and receiving coils.
In such radars, the informative parameter is the front of the reflected signal (metal is present - there is a front, metal is absent - there is no front) [10, 11]. Currently, the authors are developing the direction of identification of metals by type, by analyzing signals in the time and frequency domains with new approaches for these tasks [12, 13].
To conduct these studies, an antenna was developed that performs differential signal reception and allows you to assess the effect of the electromagnetic field on metal objects and analyze the nature of the change in the reflected signal [14-16].
The paper proposes to analyze the signal in the frequency domain and use as an informative feature in the reflected signal non-uniformity (curvature) of the envelope spectrum, which occurs after the impact of the pulsed sounding signal on the object.
The aim of this work is to develop a method of analyzing GPR signals, which is convenient for accumulating a database of known materials and for comparing the signal from an unknown object with those available in the database of standards, which allows to expand the functionality of the device. This method allows you to determine the type of metal from which the object is made.
To solve this problem, the GPR has been developed which allows you to change the parameters of the probing impulse (Fig. 1), as it is suggested to use a rectangular pulse signal and adjust the receiving antenna so that it can be recorded a transient process that occurs after the response pulse is complete.
Figure 1 - The block diagram of the developed radar system
The system consists of a receiving-transmitting antenna 1, a receiving path (block 5), a transmission path (radar block 2), a clock pulse generator 7, a processing unit on a microcontroller 6, and a storage device 3. The radar block 2 forms the pulsed signals arriving to the transmitting and receiving antenna 1 and the receiving unit 5, which is configured to receive signals from the receiving antenna 4 with the signal frequency of the radar block 2. To provide synchronization of the operation of the nodes, a clock generator is used 7. After processing signals with the radar and receiving blocks, the data is transferred to a memory of microcontroller, in which the received signals from the investigated samples are compared with the reference information stored in the memory device 3.
The low frequency signal generator 2 forms the pulse signals that arrive at the transmitter antenna 1 and emit the electromagnetic field in the test medium.
Changes in the field are recorded by the receiving antenna 4, amplified and undergoing primary processing in block 5. Synchronization between the nodes is provided by the clock pulse generator 7. After processing, the signal enters the microcontroller unit 3. In it, the signals obtained from the samples under study are compared with the ones stored in the memory unit of the block. The comparison result comes to the indicator device 6.
The input signal depends on the electrical and magnetic properties of the metals. In the theory of nondestructive testing, the influence function is used, which comprehensively shows the interaction of these characteristics on the signal:
/ o\ Vr x 2 + jP 2
91 (x, P) = ——y, , , [17]
Vr +>/X 2 + jP 2
Vr, Va, ^ - relative, absolute magnetic per-
meability of research materials and their conductivity, X = A R3, P = R^yj^VaP -generalized parameter of eddy current control [17], A -integral conversion parameter, m - angular frequency, transmitting
and receiving antennas have a radius R 3 and R ^ in accordance.
After digitization of signals and obtaining their spectra (Fig. 2-Fig. 4) on the basis of rapid Fourier transform, the identification of metals was performed by comparing the areas under the bypass [3, 11,12]. However, there are cases when the areas are close in size and the identification result may be incorrect, table 1.
Table 1
Spectral characteristics of metals [11, 12].
№ Metal Spectrum width at the level of -40dB (Hz) Area under the bypass spectrum (dB • Hz)
1 Silver 86,8% pure 6,24±0,16...26,70±0,22 535,5 ±4,6
2 Gold 90,0% pure 6,47±0,15...27,97±0,12 545,0± 3,3
3 Lead 6,64±0,29 ...28,14±0,26 547,6±4,9
4 Metal Percentage% difference between areas
5 Silver vs Gold 1,87%
6 Silver vs Lead 2,24%
7 Gold vs Lead 0,36%
To increase the reliability of identification, it is proposed to use a correlation approach [18], which consists in obtaining an additional information parameter (correlation number Kr between the central and each of
the side harmonics), which allows mutual analysis of the amplitudes of spectral components.
Consider a digitized fragment of the response signal. The following fragment, consisting of n discrete
(100-150 points) Xk, k = 1, 2,. . . n in the Fourier transform does not give a smooth envelope spectrum:
2n i
Afo & Af0+V
Afo & Af0 -N
X,
N
-kn
[18]
N-1
k =!x
n=0
Therefore, the sample x is supplemented by zeros which allows to obtain a smooth bypass spectrum.
xk + zr, k = 1,2,...n; r = 1,2
The result of the correlation analysis of the shape of the spectrum is the mutual relation with stepwise mixing from the central frequency to the lower and upper limits:
V = 1...fv N = 1...fN
The influence of the window effect [18] in the Fourier transform leads to the fact that for the same metal at different values of the input sample there will be a different step of the discrete spectral function. To offset the effect of the sample step size, the correlation result is divided by the value of the input sample ///; W + /// In this case, the correlation dependence can
be written as:
Kr =
m + n
(Af0 - A
f0 +1
)2+(Af. - Af, -1)2 +
+ iAf,~ Af,+2)2 Af,)2 +
-□ + (Af
+ (Af0 - Afw)2 + (Afn - A fM)
lfv
f0
fN ■
where A f , A
lf0 +1, Afv , AfN the amplitude of the spectrum at the central frequency in dB, at the central +1 Hz, at the upper limit of the spectrum and at the lower limit of the spectrum, respectively.
Typical spectral of signals obtained from silver 868 sample, gold 900 sample and lead are shown in Fig.2, Fig.3 and Fig.4, respectively.
5 10 15 20 25 30
Figure 2 - Typical spectrum of the reflected signal (sample material - silver 868 samples). On the abscissa the
frequency in hertz, the ordinate - the amplitude in dB
1
Figure 3 - Typical spectrum of the reflected signal (sample material - gold 900 samples). On the abscissa the
frequency in hertz, the ordinate - the amplitude in dB
Correlation numbers were calculated for these metals. They are shown in table.2.
Analysis of the results showed that the value of the correlation number can identify metals in the case when
0
the value of the area under the envelope of the spectral characteristics is close to each other. This allows to increase the reliability of metal identification by the spectral method [19-21].
-10 -
-20 -
-30
-40 -
-50 -1-'-1-1-Lj-
5 10 15 20 25 30 35
Figure 4 - Typical spectrum of the reflected signal (sample material - lead). On the abscissa the frequency in
hertz, the ordinate - the amplitude in dB
Increasing the reliability of metal distinction will reduce the number of false results and defects in production, where it is important to know the exact composition of the metal sample [22, 23].
Table 2
Comparison of metals by signals in the spectral region by correlation analysis
№ Metal Correlation numeric Kr
1 Silver 86,8% pure 4.7169±0.1117
2 Gold 90,0% pure 4.9663± 0.1736
3 Lead 4.9314±0.2631
4 Metal Percentage % difference, number Kr
5 Silver vs Gold 5,02%
6 Silver vs Lead 4,34%
7 Gold vs Lead 0,7%
The proposed correlation approach is convenient and simple in the application plan, because such calculations [24, 25] can be easily performed on modern 32bit microcontrollers [26 - 28] type STM32H7, in the manufacture of the simplest manual and automated radars that detect metal objects in dielectric environments [29, 30].
The research used metal objects of different geometric sizes and different thicknesses. It is established that for them the obtained informative features within the same metals are preserved [12, 31].
References
1. D. J. Daniels "Ground Penetrating Radar (2nd Edition)" London, UK.: Institution of Electrical Engineers, 2004, 761p. ISBN O 86341 360
2. Pravda V. I., Mrachkovskiy O. D., Abramovych A. O. Georadary, Visnik natsionalnogo universitetu «Lvivska politehnika» //Seriya Radioelektronika ta telekomunikatsiyi. - 2015. - №. 818. - C. 49-54.
3. A. O. Abramovych, O. D. Mrachkovsky, V. Yu. Furmanchuk "Dichotomic degradation of metal to black-color by spectral analysis", Bulletin of the Zhytomyr State Technological University. Ser: Technical Science, 2017, No. 1 (79), pp. 48 - 51. DOI: https://doi.org/10.26642/tn-2017-1(79)-48-51
4. A. M. Obiazi, F. I. Anyasi, O. O. Jacdonmi "Iplementing a Robust Metal Detector Utilizing the Colpitts Oscillator with Toroidal Coil", Journal of Engineering and Applied Sciences. 2010, 5(2), pp.56 -63.
5. Svatos Jakub Advanced Instrumentation for Polyharmonic Metal Detectors: Ph.D Thesis: Svatos, Jakub. - Prague, 2015. - 121p.
6. Hans A. van Sprang Fundamental parameter methods in XRF spectroscopy / A. van Sprang Hans // Advances in X-ray Analysis. - Vol.42. - 2000.
7. A. O. Abramovych, V. O. Poddubny "Application of spectral analysis for distinguishing metals by signals from eddy current converters", The East European Journal of Advanced Technology, Kharkiv,2017. Vol. 89 (No. 5), pp. 51 - 57 (Scopus) DOI: https://doi.org/10.15587/1729-4061.2017.110177
8. M. H. Jol "Ground Penetrating Radar Theory and Applications", Oxford GB.: Elsevier B.V. 2009, 574p., ISBN: 978-0-444-53348-7.
9. G.N. Scherbakov "Vybor elektromagnitnogo metoda zondirovaniya dlya poiska obektov v tolsche ukryivayuschih sred". Radiotehnika, 2005, 3, pp. 77-79 (in Russian).
10. A. O. Abramovych, V. O. Poddubny, I. S. Kashirsky "Remote vibration current analysis of metal objects composition", International scientific and technical journal "Metallophysics and the latest technologies". Kyiv: 2017. Volume 39 No. 8 pp. 1035 - 1049. (Scopus) URL: http://mfint.imp.kiev.ua/ua/abstract/v39/i08/1035.html
11. A. O. Abramovych, V. O. Poddubny, I. S. Kashirsky "Method of processing of reflected signals eddy-current pulse converters", Scientific Bulletin of Zaporizhzhya National Technical University, Radioelektronika, Informatics, Management,
Zaporizhzhia, 2017. №4, pp.7 - 14. DOI: https://doi.org/10.15588/1607-3274-2017-4-1
12. A. O. Abramovych, Y. S. Agalidi, V. O. Poddubny, " Radio engineering system identification of metals on the basis of eddy-current converters ", Scientific Bulletin of Zaporizhzhya National Technical University, Radioelektronika, Informatics, Management, Zaporizhzhia, 2020. №1, pp.7 - 17. DOI: https://doi.org/10.15588/1607-3274-2020-1-1.
13. A. O. Abramovych, V. O. Poddubny, " Eddy-Current Amplitude-Phase Based Method for Identifying Conductive (Metal) Objects", International scientific and technical journal "Metallophysics and the latest technologies". Kyiv: 2020. Volume 42 No. 8 pp. 1069 - 1085. URL: http://mfint.imp.kiev.ua/ua/abstract/v42/i08/1169.html
14. Claudio Bruschini (2002), A Multidisciplinary Analysis of Frequency Domain Metal Detectors for Humanitarian Demining, Dissertation of Doctor in Applied Sciences, Brussels, 242 p.
15. A. O. Abramovych, V. O. Poddubny "Solenoid antenna of radio engineering system of metal identification ", Bulletin of the Zhytomyr State Technological University. Ser: Technical Science, 2019, No. 1 (83), pp. 188 - 196. DOI: https://doi.org/10.26642/tn-2019-1(83)-188-196
16. Shcherbakov, G.N. (2005), «Uvelichenie predel'noi glubiny obnaruzheniya lokal'nykh ferromagnitnykh ob'ektov v tolshche provodyashchikh ukryvayushchikh sred metodom distantsionnogo parametricheskogo podmagnichivaniya», Radiotekhnika, No. 12, pp. 42-45.
17. Nerazrushayuschiy kontrol: v 5 t. T. 3. Elektromagnitnyiy kontrol: Prakt. posobie / Pod red. V. V. Suhorukova - M.: Vyissh. shk., 1992. - 312p.
18. Ajficher, Je. and Dzhervis, B. (2004), Cifrovaja obrabotka signalov. Prakticheskij podhod, Translated from English, 2nd ed., Viljams, Moskva, 992 p.
19. Kylymnyk, O.M. (2015), Pidvyshhennja tochnosti radiolokacijnogo zonduvannja pidpoverhnevyh ob'jektiv bagatochastotnymy fazovymy radiosystemamy, Dissertation of kand. teh. nauk, 05.12.17, Hmel'nyc'kyj, 164 p.
20. Diamanti, L. and Annan, P. (2013), «Characterizing the energy distribution around GPR antennas», Journal of Applied Geophysics, Vol. 99, pp. 83-90.
21. V.B. Habarov "Struktura elektromagnitnogo polya, izluchennogo podzemnyim peredatchikom s ramochnoy antennoy, s uchetom blizhney zonyi rasprostraneniya radiovoln". Radiotehnika, 2005, 3, pp. 80-83 (in Russian).
22. A. O. Abramovych, V. O. Poddubny "Rationing of signals of eddy - current converters for correct comparison of them", Bulletin of the Ternopil National Technical University. Ser: Technical sciences. Ternopil, 2017, Volume 86 (№2), pp.76 - 83.
23. A. O. Abramovych, V. O. Poddubny, O. M. Husliakov. (2017) "Eddy current device with the database, which contains different samples of metals, to identify them i n the hidden objects", Weapons and
military equipment. Kiiv, Ukraine, 15(3), pp. 51-56. doi: 10.34169/2414-0651.2017.3(15).51-56.
24. Kunts K. S. Chislennyiy analiz. Moscow, Tehnika, 1964, 389 p.
25. Gantmaher F. R. Teoriya matrits. Moscow, Nauka, 1967, 575 p.
26. Prata Stiven Yazyk programmirovaniya C++. Lektsii i uprazhneniya. Stiven Prata, Vil'yams, 2012, 1248 p. ISBN 9785-8459-1778-2
27. Glinchenko A. S. Tsifrovaya obrabotka signalov: Ucheb. posobie: V 2 ch. Ch. 2. Krasnoyarsk, IPTs KGTU, 2001, 184 p.
28. Rumshiskiy L. Z. Matematicheskaya obrabotka rezul'tatov eksperimenta. Moscow, Nauka, Glav. red. Fiz-mat. lit., 1971, 192 p.
29. R. Pernar, R. Sapina, A. Marinov, D. Vuletic, C. Matic, M. Bajic. "The relevance, strength and
likelihood of occurrence of the mine-field indicators and signatures used in the airborne and space-borne remote sensing of mine contaminated areas. New Strategies for European Remote Sensing", Oluic (ed.), 2005 Millpress, Rotterdam, ISBN 90 5966 003 X
30. Ivashov S. I., Makarenkov V. I., Razevig V. V., Sablin V. N., Sheyko A. P., Vasiliev I. A. Remote Control Mine Detection System with GPR and Metal Detector, Eight International Conference on Ground-Penetrating Radar, GPR'2000, University of Queensland, Gold Coast, Queensland, Australia, 2000, pp. 36- 39.
31. Promyishlennaya gruppa Laborant. Analizatoryi metallov i rud [Elektronniy resurs]. Rezhim dostupu: http://www.laborant.net/catalog (data zvernennya 21.04.2017), Nazva z ekrana.
