Journal of Physics: Conference Series

PAPER • OPEN ACCESS You may also like

- Numerical study on the effect of U-shaped
Detection of buried human bodies using ground- deep-buried pipe type on heat transfer
performance
penetrating radar method Yin Yuansheng, Li Chao, Yang Ruitao et
al.

- Buried Interface Sciences with X-rays and

To cite this article: Adika Bagaskara et al 2021 J. Phys.: Conf. Ser. 1876 012014 Neutrons 2010
Kenji Sakurai

- (Invited) GaN Buried Channel Normally Off

MOSHEMT: Design Optimization and
Experimental Integration on Silicon
View the article online for updates and enhancements. Substrate
Rohith Soman, Manish Sharma, Nayana
Ramesh et al.

This content was downloaded from IP address 180.244.163.138 on 09/02/2022 at 09:58

3rd International Conference on Research and Learning of Physics (ICRLP) 2020 IOP Publishing
Journal of Physics: Conference Series 1876 (2021) 012014 doi:10.1088/1742-6596/1876/1/012014

Detection of buried human bodies using ground-penetrating

radar method

Adika Bagaskara1, Abdurrahman Wafi2, Nugroho Syarif Setiawan2, Mariyanto

Mariyanto1,*
1
Department of Geophysical Engineering, Faculty of Civil Planning and Geo
Engineering, Institut Teknologi Sepuluh Nopember, Surabaya 60111, Indonesia
2
PT. Andalan Tunas Mandiri, Jakarta 13940, Indonesia

*mariyanto@geofisika.its.ac.id

Abstract. Indonesia is a country that often experiences natural disasters as it is a meeting point
of several tectonic plates. When a natural disaster occurs, most of the time the evacuation team
having some trouble finding the buried victims underground. The Ground-Penetrating Radar
(GPR) method is one of the solutions for this problem. GPR or also called georadar is a
geophysical method that is used to investigate conditions under the earth’s surface using
electromagnetic waves. This study aims to detect buried human bodies underground using the
GPR method to increase the effectiveness and efficiency of natural disaster victims by doing a
simulation measurement. It took place in one of the public cemeteries in Jakarta, Indonesia
passing two graves with corpses buried around three years before the survey. GPR
measurements were conducted using Geoscanners AKULA A9000+ Antenna GCB3070 to
detect the buried human bodies underground. The GPR data then processed using MATLAB
based program called matGPR. A series of adjustments and filters such as signal position
adjustment, remove DC, mean filter, inverse amplitude decay, remove global background, and
Karhunen-Loeve (KL) filter have been applied to the data. The result of data processing shows
amplitude contrasts which are suspected to be the buried corpses. Further research can be
conducted to investigate the buried victims from landslides, earthquakes, and other natural
disasters.

1. Introduction
Indonesia is known as a remarkable country with a lot of islands, placed between the vast Pacific and
Indian Oceans, rich with astonishing natural environments and resources along the regions. Despite its
beauty, Indonesia also lies across the Pacific Ring of Fire which is the center of some active
volcanoes, located in between tectonic plates and had intense tropical rain along the year, which
makes this country prone to devastating natural disasters such as volcano eruption, earthquake,
tsunami, and landslide [1]. As stated by BNPB, there were more than 3,397 natural disasters that
occurred in 2018 and 3,662 natural disasters occurred in 2019 [2] which threaten Indonesian people
and the environment. These natural disasters caused severe destruction and a huge number of victims.
Unfortunately, conventional methods for investigation are still often used by the evacuation team to
locate the buried victims underground. These conventional methods were taking a long time in the
process and most of the time left some buried victims unfound [3].

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
3rd International Conference on Research and Learning of Physics (ICRLP) 2020 IOP Publishing
Journal of Physics: Conference Series 1876 (2021) 012014 doi:10.1088/1742-6596/1876/1/012014

The Ground-Penetrating Radar (GPR) method is a well-known geophysical method that transmits
electromagnetic waves from antennas into the subsurface and measures the travel time between when
the radar waves are sent and when they are back to the receiver [4]. As the radar waves are transmitted
through some mediums to the buried target underground, their velocity will vary depends on the
physical characteristics of the medium in the subsurface [5]. Experiments to detect buried human
bodies using the GPR method have been done by Damiata et al. that able to detect the presence of
human skeletal inside a grave [6], and Widodo et al. that able to estimated human bodies in the
subsurface using georadar method at local graveyard [7]. All experiments mentioned before are using
the same principle, which is based on the value of the magnetic permeability and the electric
permittivity from the human objects with the environments around them [3]. This study aims to detect
buried human bodies underground using the GPR method to increase the efficiency of natural disaster
victims evacuation by doing a simulation in one of the public cemeteries.

2. Methodology

2.1. Data Acquisition

GPR measurements in this study were conducted in Jakarta, Indonesia using Geoscanners AKULA
A9000+ Antenna GCB3070 with 700 MHz of frequency. This tool was based on the GPR method
principle, which transmitting radar waves into the target medium in the subsurface and then these
waves are reflected back to the receiver. Various kinds of objects can be detected and recorded in the
GPR program from the reflection results. There are 3 acquisition lines (Line 1, 2, and 3) which are
passing 2 graves as shown in the survey acquisition design (Fig. 1) with corpses buried around three
years before the survey.

Figure 1. Survey acquisition design.

2.2. Data Processing

The acquired GPR raw data were processed using MATLAB based program called matGPR. A series
of adjustments and filters as explained below were applied to increase the interpretability of the GPR
profiles. Fig. 2 shows the flowchart of data processing.

Figure 2. Data processing flowchart.

2
3rd International Conference on Research and Learning of Physics (ICRLP) 2020 IOP Publishing
Journal of Physics: Conference Series 1876 (2021) 012014 doi:10.1088/1742-6596/1876/1/012014

2.2.1. Signal Position Adjustment. This correction is necessary because the tool used in this study did
not directly touch the ground, so all traces need to be adjusted to a time zero position before doing the
processing procedure further. This point usually is when the first negative peak of the georadar trace is
noticed [8].

2.2.2. Remove DC. DC component from each trace of the GPR profiles needs to be removed [8].

2.2.3. Mean Filter. Mean filtering is applied to decrease the spatial intensity derivatives which
occurred in the GPR profiles and generally acts as a lowpass frequency filter [8].

2.2.4. Inverse Amplitude Decay. Data amplification by the Inverse Amplitude Decay applies an
empirical gain function, which exactly compensates the mean or median attenuation observed in a 2-D
GPR radargram [8].

2.2.5. Remove Global Background. Background removal is used to remove the ringing in the GPR
data. Some noises in the data usually appear to have a similar amplitude along the same acquisition
line, this filter will also be able to remove the effect of that occurrence [9].

2.2.6. Karhunen-Loeve (KL) Filter. This filter has similarities with Fourier transforms which the
coefficients needed to transform are taken directly from the GPR data. It reconstructs the initial GPR
data with the main component which increases the lateral coherency and decreases the noise [10].

3. Results
Raw GPR profiles from the acquisition process shown in Fig. 3. These profiles gave initial
visualization from the subsurface below. Fig. 3(a) and 3(b) resembled each other because they were
taken from Line 1 and 2 which passed both graves, while Fig. 3(c) is from Line 3 which only passed
Grave 2 (see Fig. 1). The vertical scan-axis showed the two-way travel time in nanoseconds (ns) and
the horizontal axis showed the distance at which the data was recorded in meters (m). It was necessary
to process the data before doing interpretation because there were still some noises in the raw GPR
profiles.

(a) (b)

(c)
Figure 3. Raw GPR profiles of (a) Line 1, (b) Line 2, and (c) Line

3
3rd International Conference on Research and Learning of Physics (ICRLP) 2020 IOP Publishing
Journal of Physics: Conference Series 1876 (2021) 012014 doi:10.1088/1742-6596/1876/1/012014

Data adjustments and filtering was done to increase the interpretability of those raw GPR profiles.
Fig. 4(a) shows the signal position adjustment result where the signal had been positioned to time zero
because the tool used in this study did not directly touch the ground, caused a gap between the tool and
the soil. DC removal process was then applied to the GPR profile. In Fig. 4(b) The radargram contrast
seemed to become a bit lower but there was not any significant difference from the step before. The
mean spatial filter is applied to remove ringing under the target and makes the radargram in Fig. 4(c)
look smoother. Another adjustment applied to the data was the inverse amplitude decay, which
generated a significant difference in Fig. 4(d) because the signal had been strengthened. In most cases,
some noises have a similar amplitude along the same acquisition line or so-called horizontal banding.
A process of filtering called remove global background had been applied to remove the horizontal
banding and the result is shown in Fig. 4(e). The last filter applied to the data is the Karhunen-Loeve
(KL) filter. Fig. 4(f) shows the final result where the noise had been reduced and the lateral coherency
had been increased before continuing to the interpretation process.

(a) (b)

(c) (d)

(e) (f)
Figure 4. Data processing results of representative GPR profile (Line 1), involve (a) Signal position
adjustment, (b) Remove DC, (c) Mean filter, (d) Inverse amplitude decay, (e) Remove global
background, and (f) Karhunen-Loeve filter.

4. Discussion
Final radargrams from the Karhunen-Loeve filter (Fig. 5) were interpreted based on the responses
from objects in the subsurface. The amplitude contrast indicated by the blue rectangle within the travel
time range of 1.9 – 2.5 ns and the scan-axis range of 1.8 – 2.4 m in Line 1 (Fig. 5a) is suspected as the

4
3rd International Conference on Research and Learning of Physics (ICRLP) 2020 IOP Publishing
Journal of Physics: Conference Series 1876 (2021) 012014 doi:10.1088/1742-6596/1876/1/012014

location of where the body in Grave 1 is buried. There was also a higher amplitude contrast indicated
by the green rectangle within the travel time range of 1.9 – 2.7 ns and the scan-axis range of 2.8 – 3.4
m in Line 1 (Fig. 5a) which is suspected as the location of where the body in Grave 2 was buried. On
the upper part of those marked amplitude contrasts in Line 1 (Fig. 5a), there were some random
hyperbolic patterns caused by the unconsolidated rock. This soil material was not being compacted
and used to bury the body. Two amplitude contrasts were also found in Line 2 (Fig. 5b) with similar
positions of travel time and scan-axis ranges compared to Line 1 (Fig. 5a). The amplitude contrast
indicated by the green rectangle in Line 1 (Fig.5a) and Line 2 (Fig. 5b) was higher because the buried
corpse in Grave 2 might be younger than Grave 1. The body part of the younger corpse would give a
larger contrast with soil compared to the contrast of human bones from the older corpse. It was rather
difficult to interpret which amplitude contrast represents the location of the buried body in Line 3
(Fig.5c), but from the Line 1 and Line 2 results, it could be estimated that the amplitude contrast
indicated with a green rectangle was the location of buried body in Grave 2. The lower part of the
radargram in Fig. 5 shows a layering pattern from the undisturbed sand which is well consolidated
compared to the upper soil. The discontinuities on the radargram profile might be caused by the air
from the hole in the subsurface.

(a)

(b)

(c)
Figure 5. Final radargram of (a) Line 1, (b) Line 2, and (c) Line 3. The blue and green rectangles are
interpreted as the location of where the bodies are buried because of the strong amplitude contrast
presence.

5
3rd International Conference on Research and Learning of Physics (ICRLP) 2020 IOP Publishing
Journal of Physics: Conference Series 1876 (2021) 012014 doi:10.1088/1742-6596/1876/1/012014

The result from Dewi et al. is similar to the findings described in this study [3]. The significance
amplitude contrasts indicated by the rectangles (Fig. 6) are suspected as where the buried bodies are
located, while discontinuities caused by the air effect in the subsurface are indicated by red arrows [3].
It helps to prove the hypothesis that the Ground-Penetrating Radar method can be used to detect the
buried human objects underground. However, Hammon et. al mentioned that GPR frequencies need to
be above 900 MHz to give better resolution for the human body [11]. For that reason, the 700 MHz
antenna frequency used in this study might be less effective to detect the buried human bodies but it
still depends on what type of soils present in the survey location. The GPR method will perform better
in soils that are drier and sandier. It might be required to use lower antenna frequencies to detect the
buried human bodies in wetter or clay-rich soils, while higher frequencies might only be possible at
shallow depths [11].

Figure 6. Radargram result from Dewi et al. (2017) [3]

5. Conclusion
The use of the Ground-Penetrating Radar (GPR) method with 700 MHz of antenna frequency can find
high amplitude contrast anomalies which are suspected as the location of where the bodies are buried.
The amplitude contrast anomaly on Grave 2 is higher because the buried corpse might be younger,
which will give a larger contrast with soil compared to the contrast of human bones from the older
corpse in Grave 1. Further research using the GPR method can be conducted to investigate the buried
victims from landslides, earthquakes, and other natural disasters to make the evacuation process better.

Acknowledgments
We would like to thank PT. Andalan Tunas Mandiri for the ground-penetrating radar survey project
and make this study possible. We also like to express our gratitude to Institut Teknologi Sepuluh
Nopember for providing us with legal MATLAB software.

References
[1] Fatimah Y A, Murniningsih R, Setiawan A and Aman M 2019 A Smart Sustainable approach
for waste management in post-natural disaster phase IOP Conf. Ser. Mater. Sci. Eng. 674
[2] The Jakarta Post 2019 3,622 natural disasters occurred in 2019: BNPB thejakartapost.com
[3] Dewi R K, Kurniawan A, Taqwantara R F, Iskandar F M, Naufal T Z and Widodo 2017
Identification of buried victims in natural disaster with GPR method AIP Conf. Proc. 1861 1–5
[4] Conyers L B 2016 Ground-penetrating Radar for Geoarchaeology
[5] Neal A 2004 Ground-penetrating radar and its use in sedimentology: Principles, problems and
progress Earth-Science Rev.
[6] Damiata B N, Steinberg J M, Bolender D J and Zoëga G 2013 Imaging skeletal remains with
ground-penetrating radar: Comparative results over two graves from Viking Age and Medieval
churchyards on the Stóra-Seyla farm, northern Iceland J. Archaeol. Sci.
[7] Widodo W, Aditama I F, Syaifullah K, J. Mahya M and Hidayat M 2016 Detecting Buried
Human Bodies Using Ground-Penetrating Radar Earth Sci. Res.

6
3rd International Conference on Research and Learning of Physics (ICRLP) 2020 IOP Publishing
Journal of Physics: Conference Series 1876 (2021) 012014 doi:10.1088/1742-6596/1876/1/012014

[8] Szymczyk M and Szymczyk P 2014 Preprocessing of GPR data Image Process. Commun. 18
83–90
[9] Jol H 2009 Ground Penetrating Radar
[10] Zhao Y, Wu J, Xie X, Chen J and Ge S 2010 Multiple suppression in GPR image for testing
back-filled grouting within shield tunnel Proc. 13th International Conf. Gr. Penetrating Radar,
GPR 2010
[11] Hammon W S, McMechan G A, and Zeng X 2000 Forensic GPR: Finite-difference
simulations of responses from buried human remains J. Appl. Geophys.