[논문]하이브리드형 무인 항공 전자탐사시스템 자료의 분석 및 해석기술 개발

김영수; 강현우; 방민규; 설순지; 김보나

doi:10.7582/gge.2022.25.1.26

하이브리드형 무인 항공 전자탐사시스템 자료의 분석 및 해석기술 개발
Development of Data Analysis and Interpretation Methods for a Hybrid-type Unmanned Aircraft Electromagnetic System 원문보기

지구물리와 물리탐사 = Geophysics and geophysical exploration, v.25 no.1, 2022년, pp.26 - 37

김영수 (한양대학교 자원환경공학과) , 강현우 (한양대학교 자원환경공학과) , 방민규 (한양대학교 자원환경공학과) , 설순지 (한양대학교 자원환경공학과) , 김보나 (한국지질자원연구원)

초록
AI-Helper

최근의 정보기술발달에 힘입어 소형 무인 비행체를 활용한 각종 물리탐사 방법들이 제안되고 그 해석방법들에 대한 연구가 소개되고 있다. 이 연구에서는 한국지질자원연구원에서 개발 중인 송수신 분리형 무인 항공 전자탐사 장비를 소개하고 획득한 자료의 타당성 검증을 위해 수행된 시험자료를 분석하여 해석하는 방법을 제안하는 연구를 수행하였다. 특히, 수신기가 드론에 매달린 채로 탐사가 수행되기 때문에 발생되는 흔들림 성분의 영향을 고찰하고 회전변환을 이용하여 보정하였다. 한편, 비행체에 의한 탐사는 송수신기 간의 거리, 고도 등 여러 탐사 변수들이 실시간으로 변하게 되고 획득한 자료는 지상 탐사보다 더 많은 잡음을 포함하게 되어 전통적인 해석방법으로의 해석에 많은 어려움이 따른다. 따라서, 이 연구에서는 획득한 전자탐사자료를 이용하여 빠르게 겉보기 비저항을 예측할 수 있는 순환 인공 신경망 모델을 구축하였으며, 현장자료의 분석을 통해 얻어진 잡음들을 수치모델링을 통해 생성한 학습자료에 포함시켜 잡음이 포함된 자료의 예측성능을 향상시켰다. 학습된 순환 신경망 모델을 시험탐사 현장자료에 적용시킨 결과 지상탐사 및 전기비저항 탐사 결과와 유사한 겉보기 비저항을 예측함을 확인하였다.

Abstract ▼ AI-Helper

Recently, multiple methods using small aircraft for geophysical exploration have been suggested as a result of the development of information and communication technology. In this study, we introduce the hybrid unmanned aircraft electromagnetic system of the Korea Institute of Geosciences and Mineral resources, which is under development. Additionally, data processing and interpretation methods are suggested via the analysis of datasets obtained using the system under development to verify the system. Because the system uses a three-component receiver hanging from a drone, the effects of rotation on the obtained data are significant and were therefore corrected using a rotation matrix. During the survey, the heights of the source and the receiver and their offsets vary in real time and the measured data are contaminated with noise. The noise makes it difficult to interpret the data using the conventional method. Therefore, we developed a recurrent neural network (RNN) model to enable rapid predictions of the apparent resistivity using magnetic field data. Field data noise is included in the training datasets of the RNN model to improve its performance on noise-contaminated field data. Compared with the results of the electrical resistivity survey, the trained RNN model predicted similar apparent resistivities for the test field dataset.

주제어

표/그림 (21)

그림 Fig. 1. Photo of the hybrid-type unmanned aircraft electromagnetic system. (a) Korea Institute of Geosciences and Mineral resources airship with the mounted source loop indicated by the yellow line. (b) Three-component receiver with the Inertial Measurement Unit (IMU) indicated by the orange circle. (c) Drone with the threecomponent receiver indicated by the red circle.
그림 Fig. 2. Amplitude data for the three-component magnetic fields with time recorded at the same surface position with an offset of 50 m from the source. The red, blue, and green lines represent H_x, H_y, and H_z, respectively.
그림 Fig. 3. Survey geometry of the surface field test. The source and receiver locations for the data acquisition are plotted on the aerial photo of the test site.
그림 Fig. 4. (a) Variation in the amplitude ratio of the vertical to horizontal magnetic fields with time. (b) Histogram of the amplitude ratios picked from panel (a). The mean value of the amplitude ratios is indicated by the red dot on the x-axis. (c) Procedure to identify the apparent resistivity from the ratio of the vertical to horizontal magnetic components. The green graph represents the ratio of the vertical to horizontal magnetic components as a function of the halfspace resistivity.
그림 Fig. 5. Schematic diagram of the survey geometry for the numerical modeling. The horizontal offset between the source and the receiver varies from 70 m to 100 m. The electrical resistivity of the halfspace model changes from 10 ohm-m to 500 ohm-m.
그림 Fig. 6. Comparison of the apparent resistivity maps from the direct current (DC) and electromagnetic (EM) surveys. (a) Apparent resistivity map obtained by the DC resistivity survey. (b) Prediction of the apparent resistivity map calculated from the EM data.
그림 Fig. 7. Receiver locations for the data acquisition plotted on an aerial photo of the test site for the drone flight. The three-component receiver is moved by the drone while the source is fixed at the surface.
표 Table 1. Mean and standard deviation values of the apparent resistivity at each receiver position.
그림 Fig. 8. Schematic diagrams of the three rotation components: (a) roll, (b) pitch, and (c) yaw.
그림 Fig. 9. Measured rotation component data according to the measuring time. The orange line indicates the roll component, and the purple line indicates the pitch component. The green line near 60 s indicates the change in the flight direction. Panel (a) plots data from the first flight, and panel (b) plots data from the second flight.
그림 Fig. 10. Variation in the horizontal to vertical magnetic field ratio according to the changes in the rotation components. The effects of rolling and pitching are compared for the reference height of the receiver (5 m), and response variations are shown for height deviations of ±0.75 m from the reference height.
그림 Fig. 11. Calculated ratio of the vertical to horizontal magnetic fields according to the offset from the source for the first flight. The blue line represents the original data, and the black line represents the data corrected using the rotation matrix. The green line at a 100 m offset indicates the change in the flight direction.
그림 Fig. 12. Workflow for predicting the apparent resistivity using the pretrained recurrent neural network (RNN) model.
그림 Fig. 13. Comparison of the modified one-dimensional modeling results plotted as a red line with the results from the threedimensional modeling plotted as a blue line when the half-space resistivity is 100 ohm-m and the height is 5 m.
그림 Fig. 14. Conceptual diagram of the RNN.
그림 Fig. 15. Ratio of the horizontal to vertical components according to the resistivity for each offset when the height is 5 m. Panel (a) shows the vertical component over the horizontal components, and panel (b) shows the horizontal components over the vertical component.
그림 Fig. 16. Schematic diagram of the survey geometry for the numerical modeling. The horizontal offset between the source and the receiver varies from 60 m to 100 m, and the height of the receiver varies from 3 m to 6.5 m. The electrical resistivity of the half-space model changes from 1 ohm-m to 1000 ohm-m.
그림 Fig. 17. Verification of the performance of the trained RNN model. The dashed black lines represent the true resistivity values of the test data, and the straight colored lines represent the predicted apparent resistivity values given by the trained RNN model for different offsets.
그림 Fig. 18. Analysis of the noise patterns for the amplitude ratio and the height of the receiver. (a) Field data contaminated with noise (orange lines), with the major trend of the data shown by the blue lines. (b) Histogram according to each variance between the data and the major trend (orange blocks) and the best fitted Gaussian distribution (blue lines).
그림 Fig. 19. Verification of the performance of the trained RNN model. The dashed black lines represent the true resistivity values of the noise-added test data, and the solid-colored lines represent the predicted apparent resistivity values using two different RNN models trained with different input data for different offsets. (a) Results predicted by the original RNN model. (b) Results predicted by the RNN model trained with the noise-added dataset.
그림 Fig. 20. Prediction results for field datasets obtained via two roundtrip flights using the RNN model pretrained with the noise-added dataset.

참고문헌 (18)

Auken, E., Christiansen, A. V., Westergaard, J. H., Kirkegaard, C., Foged, N., and Viezzoli, A., 2009, An integrated processing scheme for high-resolution airborne electromagnetic surveys, the SkyTEM system, Explor. Geophys., 40(2), 184-192, https://doi.org/10.1071/EG08128

상세보기
Ball, V., Blangy, J. P., Schiott, C., and Chaveste, A., 2014, Relative rock physics, TLE, 33(3), 276-286, https://doi.org/10.1190/tle33030276.1

상세보기
Bang, E. S., Son, J. S., Kang, W., Yi, H., Kim, C., Lee, C. W., Kim, B., Hwang, S., No, S., Son, Y., and Cho, S. J., 2020, Development of a Drone Platform by KIGAM for Geological Surveys and Mineral Resource Exploration, Geophys. and Geophys. Explor., 23(3), 141-148 (in Korean with English abstract), https://doi.org/10.7582/GGE.2020.23.3.00141

원문보기 상세보기
Bang, M., Oh, S., Seol, S. J., Lee, K. H., and Cho, S. J., 2018, Theoretical Research for Unmanned Aircraft Electromagnetic Survey: Electromagnetic field calculation and analysis by arbitrary shaped transmitter-loop, Geophys. and Geophys. Explor., 21(3), 150-161 (in Korean with English abstract), https://doi.org/10.7582/GGE.2018.21.3.150

원문보기 상세보기
Bang, M., Oh, S., Noh, K., Seol, S. J., and Byun, J., 2021, Imaging subsurface orebodies with airborne electromagnetic data using a recurrent neural network, Geophysics, 86(6), 407-419, https://doi.org/10.1190/geo2020-0871.1

상세보기
Cho, I. K., Kim, R. Y., and Yi, M. J., 2015, One-dimensional modeling of airborne transient electromagnetic using a long grounded-wire source, Geophys. and Geophys. Explor., 18(4), 216-222 (in Korean with English abstract), https://doi.org/10.7582/GGE.2015.18.4.216

원문보기 상세보기
Cho, I. K., Jang, J. H., Rim, H. R., and Yi, M. J., 2017, Laterally constrained inversion of GREATEM data, Geophys. and Geophys. Explor., 20(1), 33-42 (in Korean with English abstract), https://doi.org/10.7582/GGE.2017.20.1.033

원문보기 상세보기
Chung, Y., Son, J. S., Lee, T. J., Kim, H. J., and Shin, C., 2014, Three-dimensional modelling of controlled-source electromagnetic surveys using an edge finite-element method with a direct solver, Geophys. Prospect., 62(6), 1468-1483, https://doi.org/10.1111/1365-2478.12132

상세보기
Dobrovolskiy, A., Brass, Kristaps., 2021, GPR on a drone-Practical experience gained in the past three years, First Internat. Mtg for Applied Geoscience and Energy., Soc. Expl. Geophys., Expanded Abstracts, 3048-3052, https://doi.org/10.1190/segam2021-3591560.1
Fournier, D., Kang, S., McMillan, M. S., and Oldenburg, D. W., 2017, Inversion of airborne geophysics over the DO-27/DO18 kimberlites - Part 2: Electromagnetics. Interpretation, 5(3), T313-T325, https://doi.org/10.1190/INT-2016-0140.1

상세보기
Giles, C. L., Lawrence, S., and Tsoi, A. C., 2001, Noisy time series prediction using recurrent neural networks and grammatical inference, Mach. Learn., 44(1), 161-183, doi: 10.1023/A:1010884214864
Hammack, R., Veloski, G., Schlagenhauf, M., Lowe, R., Zorn, A., and Wylie, L., 2020, Using drone-mounted geophysical sensors to map legacy oil and gas infrastructure, URTEC 2020, 3517-3528, https://doi.org/10.15530/urtec-2020-2876
Kang, H., Bang, M., Kim, Y. S., and Seol, S. J., 2021, Analysis of test data obtained from hybrid-type unmanned aerial electromagnetic system, 2021 Spring Conference of KSEG, 32 (in Korean).
Kim, H. J., Song, Y., and Lee, K. H., 1997, High-frequency electromagnetic inversion for a dispersive layered earth, J. Geomag. Geoelec., 49(11-12), 1439-1450, https://doi.org/10.5636/jgg.49.1439

상세보기
Noh, K., Yoon, D., and Byun, J., 2020, Imaging subsurface resistivity structure from airborne electromagnetic induction data using deep neural network, Explor. Geophys., 51(2), 214-220, https://doi.org/10.1080/08123985.2019.1668240

상세보기
Oh, S., Noh, K., Yoon, D., Seol, S. J., and Byun, J., 2018, Salt delineation from electromagnetic data using convolutional neural networks, IEEE Geosci. Remote. Sens. Lett., 16(4), 519-523, https://doi.org/10.1109/LGRS.2018.2877155

상세보기
Pfaffhuber, A., Anschuetz, H., Rafezi, H., Novo, A., Hassani, F. P., Sorensen, K. I., Liu, W., Chen, R., Wu, H., Qiu, J., Yao, H., Shen, R., Ren, Q., Chang, F., Zeng, P., Luo, W., Hodges, G., Garrie, D., Christensen, C., Legault, J. M., Toop, D., Oldenborger, G. A., Plastow, G., Bournas, N., Han, Z., Orta, M., Fage, I., Hodges, G., Chen, T., Cox, L., Endo, M., Zhdanov, M., Paine, J. G., Collins, E. W., Costard, L., Baranwal, V., Ronning, J. S., Solberg, I.-L., Dalsegg, E., Tonnesen, J. F., and Ahmed, S., 2015, Airborne geophysics, remote sensing, UAV (drone)-based surveys and mining geophysics, Symp. Appl. Geophyics Eng. Environ. Probl., 40-53, https://doi.org/10.4133/SAGEEP.28-005
Puzyrev, V., 2019, Deep learning electromagnetic inversion with convolutional neural networks, Geophys. J. Int., 218(2), 817-832, https://doi.org/10.1093/gji/ggz204

상세보기

저자의 다른 논문 :

표제어: PCR

동의어: Packet Collision Rate

용어 설명 출처 목록 (6)

용어 설명: PCR은 세균 특이성이 있는 primer를 이용하여 적은 수의 세균이 있을지라도 쉽게 검출할 수 있는 유용한 방법이며, 이를 이용하여 구강 내 치면세균막이나 타액에서 직접 세균을 검출할 수 있게 되었다[8].

내보내기 구분	파일저장 인쇄 메일전송
구성항목	기본정보 상세정보 관리번호, 논문명, 저널/프로시딩명, 저자 , 발행년, 권, 호, 시작페이지, 끝페이지, 발행기관 관리번호, 논문명, 대등논문명, 저자 , 저널/프로시딩명, 발행기관, 발행년, 발행언어, 권, 호, 시작페이지, 끝페이지, ISBN, ISSN, 주제분야, 키워드, 초록(한글), 초록(영문), 저자(소속기관)
저장형식	Text(ASCII format) Excel format RefWorks Direct Export RIS format (for Reference Manager, ProCite, EndNote), Scholar's Aids, Mendeley
메일정보	받는사람 (필수) @ 보내는사람 (선택) @ 제목 내용 KISTI 검색결과 이메일 서비스
안내	총 건의 자료가 검색되었습니다. 다운받으실 자료의 인덱스를 입력하세요. (1-10,000) 검색결과의 순서대로 최대 10,000건 까지 다운로드가 가능합니다. 데이타가 많을 경우 속도가 느려질 수 있습니다.(최대 2~3분 소요) 다운로드 파일은 UTF-8 형태로 저장됩니다. 파일의 내용이 제대로 보이지 않을실 때는 웹브라우저 상단의 보기 -> 인코딩 -> 자동선택 여부를 확인하십시오. ~ Text(ASCII format) Excel format

연합인증