Development and Efficacy of Laser Bird Repeller with PTZ Camera

and Caffe Framework

SUBONG PYON1,*, JONGSU SO1, JUNHYOK PAK1, JINZU SO1, SUNIL KIM1

1School of metallurgical Engineering, Kim Chaek University of Technology, Pyongyang 999093,

DPR OF KOREA

Abstract: - Many researches have been conducted to prevent bird damage not only on farms, orchards, fish

farms and airports but also on industrial and urban environments. In this paper, we describe the construction

and effectiveness of a laser bird repellent system that scaring it away as soon as the bird settles in a large area.

The object classification technique using the Caffe framework detects the bird in real-time captured images of a

PTZ camera over a large area, and when the bird sinks into the area, the PTZ camera combining the laser beam

generator is steered to the target bird, thereby scaring the bird away by attacking it by the laser beam.

Key-Words: - Bird repellent, laser beam, Caffe framework, PTZ camera, Image Classification, Object Detection

Received: March 9, 2024. Revised: August 7, 2024. Accepted: September 11, 2024. Published: October 17, 2024.

1. Introduction

Bird damage not only farms, orchards, fish farms,

airports [1], [2], but also industrial and urban

environments [3], [4]. Common methods for

preventing bird damage include the use of

scarecrows [5], [6], the use of chemicals [6], and the

use of nets [7]. In recent years, studies that use

ultrasound [11], [12], studies that use sound waves

[9], [10], and studies that use laser [13] have also

been carried out to repel bird.

Most birds are known to have a very sensitive

response to light. [14]

Therefore, bird repellent using laser beams has

been widely used in both manual and automatic

modes as a method of chase bird by influencing the

vision of bird in a wide area using the characteristics

of lasers. A manual laser bird repeller is a device

that chases birds by manually controlling the laser

beam collection point of the repeller to the bird

position when a person is monitoring a certain area

and finding the bird.

An automatic laser bird repeller is a system that

allows birds around a track to fly in surprise by

repeatedly moving the laser beam spot along a

predetermined trajectory in a certain area. [14]

Because the automatic laser bird repeller repeats

the operation of illuminating the laser beam along a

certain trajectory, with or without the presence of a

bird, it is judged difficult to extrude it from the

invading area immediately after the bird has

invaded.

This may be one factor that prevents the

automatic laser repeller from having a high bird

repelling effect.

In addition, the long operation time of the laser

beam generator and the control device can have a

strong effect on the compliance degree of the bird

and the lifetime of the devices.

This makes it possible to realize that a bird

repelling approach is necessary to minimize the

operation time of the laser beam generator and the

control device and to scare the bird away that have

invaded the area as quickly as possible.

Recently, methods to automatically track and

classify birds using deep learning [15],[16],17]

together with methods to track birds’ movements

using cameras[18],[19],[20] have been introduced.

The objective of this study was to construct an

intelligent system and measure effect of it to target

and expel the invaded bird with a laser beam by

running a camera controlling device and a laser

beam generator instantly when the bird invaded and

landed within an area that requires bird repelling.

First, the target area is monitored in real time by

a PTZ camera and the intrusion (sinking) position of

the bird object is determined using the object

classification technique by the Caffe framework.

The control device of the PTZ camera coupled

with the laser beam generator is controlled to the

bird position and the bird is expelled from the target

area by outputting the laser beam.

Such a system can increase the bird repellent

performance and efficiency by minimizing the

operation time of the laser beam generator and the

control device of camera, thereby increasing the

service life of the system and reducing the

compliance of the bird, as well as immediately

scaring a wide range of birds.

Section 1 summarizes the architecture and

structure of the system.

Section 2 presents the devices that make up the

system and their combinations, and Section 3

EARTH SCIENCES AND HUMAN CONSTRUCTIONS

DOI: 10.37394/232024.2024.4.3

Subong Pyon, Jongsu So, Junhyok Pak,

Jinzu So, Sunil Kim

E-ISSN: 2944-9006

Volume 4, 2024

summarizes the bird classification using the Caffe

framework.

Section 4 presents the architecture of the

software that comprises the system and Section 5

presents the experimental results.

2. Formation of system

The system's configuration consists of five

components (Fig.1).

Fig 1: Construction of System

2.1 Image entry unit

Capturing and displaying the image in real time on

the PTZ camera.

The Image entry unit consists of a camera, a

network hub, a network line, a control computer,

and image capture and display software.

Image capture and display use an SDK that is

supported by the camera manufacturer.

In addition, due to the different types of cameras,

the obtained images are color-transformed in real

time into RGB images and sent to the following

image processing unit.

2.2 Image processing unit

Processing the acquired real-time image to obtain

the location of the bird.

The image processing unit consists of a control

computer and image processing software.

In the captured real-time image, motion detection is

performed to obtain all motion objects and to

classify each object into a bird or not.

If the moving object is judged to be a bird, the

PTZ position of the bird becomes the input

parameter of the following control panel movement

unit.

2.3 Control panel movement unit

Places the focus center of the camera in the center of

the target bird.

The focus of the camera is placed at the center of

the bird by controlling the camera by software to the

PTZ position of the bird.

And according to the mode of attack, the camera

control panel is further controlled.

The camera and the laser beam generator are

integrated, and the focus center of the camera and

the laser scanning center are aligned.

2.4 Laser beam control unit

If a bird is detected in the image processing unit,

laser attack according to the mode of attack is

performed with the movement of the camera control

panel.

The laser beam control unit consists of a laser

beam control panel, a laser beam control program,

and a laser beam generator.

When the laser attack signal and the attack mode are

transmitted to the laser beam controller, the

controller initiates the attack by switching on the

power of the connected laser beam generator and

stops the generation of the laser beam by switching

off the power of the laser beam generator when the

attack termination signal is transmitted.

Depending on the mode of attack, the attack

effect can be enhanced by selectively transmitting or

simultaneously transmitting the attack signal to a

laser beam generator of blue (420 nm), green (532

nm), or red (650 nm).

2.5 User interface

It is an interface program that enables RS-485

communication with the computer and the bird

detector and network communication with the

camera as user interface of the intelligent laser bird

repellent system, and to manually and automatically

browse the recording data, view the statistical data,

and set various settings together with the control of

the laser beam attack and camera control unit.

3. Devices

The system includes a PTZ camera, a laser beam

controller, a laser beam generator, a coupling sleeve,

and a control computer for real-time image

acquisition of the bird monitoring area.

3.1 PTZ camera

HIKVISION (type number DS-2DC6120IY-A,

Firmware V5.3.22 build 181120, resolution 1280 x

720) was used as the main device for the Image

entry unit.

The camera is outdoor, capable of PTZ control

through a computer network, supports the ONVIF

protocol, and has a 3D Position (3D Position)

function.

3.2 Laser beam controller

It is a device for controlling the start and stop of

operation of each laser beam generator.

ARM32-bit CortexTM-M3 CPU Core was used as a

central processing chip to ensure good performance.

EARTH SCIENCES AND HUMAN CONSTRUCTIONS

DOI: 10.37394/232024.2024.4.3

Subong Pyon, Jongsu So, Junhyok Pak,

Jinzu So, Sunil Kim

E-ISSN: 2944-9006

Volume 4, 2024

The circuit of this device consists of the main

circuit, the power circuit, the communication circuit,

and the gain control circuit (Fig.2).

Fig 2: Configuration of laser beam controller

The input power is AC 220 V and is stabilized by

DC 12 V and 5 V in the supply circuit to supply the

main circuit and the peripheral circuit.

The resulting signal, which consists of an ARM

chip and its driving circuits, operates on this chip to

control the laser beams through the amplifier circuit.

The communication circuit part used RS485 or

Ethernet communication mode as a circuit for

transmitting and receiving data between the chip

and the computer and between the chip and the

camera.

3.3 Laser beam generators

The laser beam generator (1000 mW Class-III

Laser) had 1000 mW power of red, green and blue

and was fixed to the coupling sleeve and coupled to

the inside of the PTZ camera.

At this time, the overnight lighting LEDs inside

the PTZ camera were first separated from the fixture

and the coupling sleeve was fixed there.

3.4 Connecting bush

The laser beam generator is fixed to the camera and

is a device for matching the focus of the camera

with the laser scanning center.

A laser beam generator of three colors, red, blue

and green, was installed in two each.

The laser beam generator and the connecting bush

for mounting on the camera are shown in Fig. 3.

After all mounting, the bolts of the connecting

bush were adjusted so that the focus of the laser

beams was matched to the viewing image focus of

the camera.

Fig 3: Laser beam generator and connecting

bush

3.5 A computer for control

In the computer for control, a bird repellent program

is executed, such as processing images captured in

real time through a local area network (Ethernet)

from an IP camera, classifying birds using the Caffe

model, and transmitting attack start and attack finish

signals to the laser beam controller using RS-485.

The operating system is Windows 7-x64, the

computer processor is Intel Core (TM) i5-3340 M

CPU @ 2.70 GHz, and RAM 4.00 GB.

4. Moving Object Detection and Bird

Classification

4.1 Moving Object Detection

First, we detected moving objects in the surveillance

area image captured in real time from the camera.

Motion detection was used with the Background

Modeling Detector (Aforge.Net Framework 2.2.5

version).

The difference threshold (DifferenceThreshold) of

the detector was set to standard value 10 and the

background update frame value

(FramesPerBackgroundUpdate) was set to standard

value 5.

4.2 Train data gathering

First, training data were collected for bird

classification using Caffe.

The training data were constructed by storing the

bird candidates obtained by motion detection

described above as files in real time and classifying

them as bird and no-bird.

When selecting bird images, flying birds were not

selected as bird candidates, and only sedentary birds

were selected as learning objects.

The bird images collected for training were

31,000 (23,000 magpies, 8,000 crows), and the no-

bird images were 30,500 color BMP files, which

were stored in different folders (bird images in birds

folder and other images in others folder).

EARTH SCIENCES AND HUMAN CONSTRUCTIONS

DOI: 10.37394/232024.2024.4.3

Subong Pyon, Jongsu So, Junhyok Pak,

Jinzu So, Sunil Kim

E-ISSN: 2944-9006

Volume 4, 2024

After deep learning, the data to be used for testing

are stored in the Test folder (3,000 bird images,

3,000 no-bird images).

4.3 Bird Training

For deep learning, preprocessing of the training data

was performed first.

For preprocessing, gray-scale and image size

conversion are used, and image size conversion is

set to 50*41. After pre-processing, deep learning is

performed, when the Mobile Network 2.0 version

(mobilenet_v2) is used.

The mobile network version 2.0 is composed of four

overlay layers (Convolution Layer), four selection

layers (Pooling Layer), and two fully connected

layers (Fully Connected Layer) with a gray-tone

image of size 198 x 161 as input to the neural

network and the position of five reference points as

output.

In Solver.prototxt, change only the maximum

number of iterations (max_iter) to 500,000 times,

and in Train.prototxt, change the input layer to

ImageData.

layer {

name: "data"

type: "ImageData"

top: "data"

top: "label"

include {

phase: TRAIN

}

transform_param{

scale: 0.0078125 # 128

mirror: true

}

image_data_param{

source: "./prototxt/trainval2.txt"

#batch_size: 32

batch_size: 16

shuffle: 1

is_color: false

new_height: 41

new_width: 50

}

Since our system is a Binary Classification (Binary

Classification) problem that perform laser attack if it

is a bird and does not perform laser attack if it is not

a bird, the number of the last output network

(num_output) is set to 2.

Deep learning is done by re-learning (finetuning)

the default model of Mobile Network version 2.0.

After the training was completed, the test data were

tested to obtain a classification accuracy of 95.2%.

4.4 Bird Classification

We used the newly constructed network model to

perform bird classification in real time.

The resulting images of motion detection from the

real-time input image from the camera were fed into

the input of the bird classification library using the

newly constructed network model, and the outputted

classification accuracy was judged to be bird if it

was more than 0.5, laser attack was performed, or

not.

5. Software

The software was written for the 64-bit operating

system using the C-Sharp development language of

Microsoft Visual Studio 2015.

Interface for camera SDK. HCNetSdkCom, SDK

of HIKVISION camera, is used. In the

CHCNetSDK.cs file, the CHCNetSDK class that

maps all internal functions and structures of the

SDK library in C-Sharp language is defined and

used. The use of all functions of the camera, such as

camera connection, setting, real-time image

collection and display, and camera control, is

conducted through this face-to-face.

Interface for image process library. A library

including image processing modules necessary for

the system was prepared and used. The function to

convert the YUV12 image acquired by the camera

to RGB image (GIP_CovertToRGB function),

initialization of Caffe library (GIP_CaffeInit) and

bird classification (GIP_CaffeClassify) functions

have been implemented, and real-time image

processing is performed through this interface.

Function of capturing and display from camera.

The image of the PTZ camera is captured in real

time and displayed by dividing the screen according

to the number of cameras.

Function of control panel manual control and

camera setting. The camera SDK is used to

manually control the PTZ movement of the camera

and includes the camera setting function, such as the

movement speed and the basic positioning.

Function of auto and manual attack. In the

manual attack, if any position in the real-time

captured display image is selected as a rectangle, the

camera moves the PTZ to that position, allowing the

laser beam generation to start operation as well.

At the end of the attack, the laser beam is returned

to the default position, and the laser beam

generation is also stopped.

In an automatic attack, if a bird classification is

performed in real time and judged to be a bird, i.e.,

the bird classification result has a similarity value of

0.5 or more, then PTZ moves the camera to the

scene position, initiates the generation of the laser

EARTH SCIENCES AND HUMAN CONSTRUCTIONS

DOI: 10.37394/232024.2024.4.3

Subong Pyon, Jongsu So, Junhyok Pak,

Jinzu So, Sunil Kim

E-ISSN: 2944-9006

Volume 4, 2024

beam, returns to the default position after the attack

according to the attack mode, and stops the

generation of the laser beam.

Function of attack history view. The ability to

browse video data recorded on a hard disk during an

attack on a bird by recording time.

Function of statistics view. It is a function of

browsing the statistics of attack on birds on a bar

chart.

The start and end dates were selected so that the

number of bird invasions and the number of attack

successes during the period were displayed as

statistics.

Function of system configuration. The system

setup consists of the network control functions such

as IP address and user and password setting of the

cameras and COM communication setup functions

such as COM port and communication band, video

storage location setting, and laser attack mode

setting functions.

6. Results

6.1 Test condition

For the experiment, a square airspace area of 200m

x 200m in Moran Hill, Pyongyang, where magpies

and crows flew a lot, was selected, and the area was

divided in half and a total of two our bird repellers

were installed at 20m height of pine trees in a

position to monitor each area.

There are 28 pine trees and 520 m2 of pond in the

area of the open area, while the rest are lawn areas.

Depending on the distance from the camera (50

m, 100 m, and 200 m), a red flag was placed at both

ends of the zone to make the distance visible on the

screen.

The experiment was conducted from October

2019 to August 2023 throughout the spring to

winter, during which 31,308 birds (21,142 magpies,

8046 crows, and 120 others) were observed to

invade the area.

In addition, bird infestation was observed from 5

to 19 a.m. and almost no other time.

And the experiment was carried out by changing the

laser color from blue, red and green every day.

6.2 Classification result

For the results of deep learning, an evaluation of the

bird classification performance was performed.

. For 10308 magpies and crows, positive (False

Positive Rate) and negative (False Negative Rate)

sampling rates were calculated.

Here, the positive sample rate is the rate

estimated by fitting a moving object with no

magpies or crows, and the negative sample rate is

the rate estimated by not being magpies or crows for

a moving object.

In our study, FPR = 3.8% and FNR = 1.2% were

calculated.

6.3 Effective rate by Laser color and attack

distance

Experiments were carried out to evaluate the

response of laser color and distance from 5 a.m. to

19 p.m. when the attack mode was changed to red,

green and blue in the laser control unit according to

the laser color and distance, with bird infestations

such as magpies and crows being the most frequent.

If a bird subjected to laser attack evades the

invasive area within 3 s, it is considered to have an

attack effect, or it is considered to have no attack

effect if not.

For the total number of birds measured, the

percentage of the number of birds that had an attack

effect was evaluated as responsive.

The distance-dependent flags were estimated to

be 50 m distance when the invading bird were near

40-75 m, 100 m distance when near 75 m to 150 m,

and 200 m distance from 150 m to 200 m.

The total measured data were calculated by

arithmetic averaging because the magpies and crows

were similar in responsiveness and the illuminance

varied with time over the seasons.

The figures below show the reactivity of the

algae at distances of 50 m, 100 m and 200 m,

respectively (Fig. 4, Fig. 5, Fig. 6).

Fig 4: Response effect rate according to the laser

color (50m)

EARTH SCIENCES AND HUMAN CONSTRUCTIONS

DOI: 10.37394/232024.2024.4.3

Subong Pyon, Jongsu So, Junhyok Pak,

Jinzu So, Sunil Kim

E-ISSN: 2944-9006

Volume 4, 2024

Fig 5: Response effect rate according to the laser

color (100m)

Fig 6: Response effect rate according to the laser

color (200m)

Graph shows that the reactivity with Laser Color

depends on the illuminance.

It can be seen that the response of the bird is the

lowest, especially at noon, when the illuminance is

greatest, and the response is large in the morning

and evening when the illuminance is relatively

small.

It can also be seen that the green laser has a

relatively large bird repelling effect and a low

reactivity with distance.

6.4 Shirking Time of bird

We measured the time it took for a bird to start to

run out of surprise from the time it was perceived by

the camera to enter the survey area.

The magpies or crows flew in astonishment as

soon as a circular beam of light appeared, as the

laser beam descended into the field of the intrusion

area.

With the ranging accuracy of the laser beam

generator and the slight movement of the bird, the

effect is different, but the evacuation time is

estimated to be about 0.5 s to 3 s.

This was attributed to the inability of the bird to

perform its desired behavior in the descending

position.

7. Conclusion

Considering the bird classification performance and

reactivity, the bird repelling rate of the system is

estimated to be 93% in the morning and evening and

7% in the day time within the distance 200 m region

when using a 1000 mW green laser.

And the prevention area of one bird repeller is

estimated at about 2000 m2 in daylight and 25 000

m2 in morning and evening.

The operating time of the laser beam generator

and the camera control panel is about 5 to 7 seconds

per cycle starting the attack and returning to its

position, and the operating time of the day is equal

to the operating time per cycle multiplied by the

number of birds detected.

It can be seen that the service life is much longer

and the compliance of the algae is as small as that of

the scanning laser algae detector.

Intelligent laser bird repeller using PTZ cameras

and Caffe frameworks have high bird repellent

effect due to their long service life, large area of

protection, and prompt repelling of bird infestations,

and can be widely used in farms, aquaculture farms,

orchards, airports, etc.

Acknowledgement:

We would like to thank Mr. Kwanghwi Kim, who

taught us about the need for bird control and pushed

us into this research.

References:

[1] Author, Title of the Paper, International

Journal of Science and Technology, Vol.X,

No.X, 200X, pp. XXX-XXX.

[2] Author, Title of the Book, Publishing House,

200X.

[3] Anderson, A., Lindell, C., Moxcey, K. M.,

Siemer, W., Linz, G. M., Curtis, P., Steensma,

K. “Bird damage to select fruit crops: the cost

of damage and the benefits of control in five

states,” Crop Protection, vol. 52, pp.103-109,

2013.

[4] York D. L., Cummings J. L., Engeman R. M.,

and Davis Jr. J. E. “Evaluation of Flight

Control and Mesurol as repellents to reduce

horned lark (Eremophila alpestris) damage to

lettuce seedlings,” Crop Protection, vol. 19,

no.3, pp.201-203, 2000.

[5] Y. Yang, Y. Zhou, Z. Feng, and K. Wu,

“Making the Case for Parks: Construction of an

EARTH SCIENCES AND HUMAN CONSTRUCTIONS

DOI: 10.37394/232024.2024.4.3

Subong Pyon, Jongsu So, Junhyok Pak,

Jinzu So, Sunil Kim

E-ISSN: 2944-9006

Volume 4, 2024

Ecological Network of Urban Parks Based on

Birds,” Land, vol. 11, no. 8, Aug. 2022,doi:

10.3390/LAND11081144.

[6] W. Xu et al., “Bird Communities Vary under

Different Urbanization Types—A Case Study

in Mountain Parks of Fuzhou, China,”

Diversity, vol. 14, no. 7, Jul. 2022,

doi:10.3390/D14070555.

[7] Summers R. W., “The effect of scarers on the

presence of starlings (Sturnus vulgaris) in

cherry orchards,” Crop Protection, vol.4, no.4,

pp.520- 528, 1985.

[8] Vickery J. A. and Summers R. W., “Cost-

effectiveness of scaring brent geese Branta b.

bernicla from fields of arable crops by a human

bird scarer,” Crop Protection, vol. 11, no.5,

pp.480-484, 1992.

[9] Avery M. L., Werner S. J., Cummings J. L.,

Humphrey J. S., Milleson M. P., Carlson C. J.,

Primus T. M. and Goodall M. J. ,”Caffeine for

reducing bird damage to newly seeded rice,”

Crop Protection, vol. 24, pp.651-657, 2005.

[10] Bruggers R. L. and Ruelle P. “ Efficacy of nets

and fibers for protecting crops from grain-

eating birds in Africa,” Crop Protection, vol. 1,

no.1, pp.55-65, 1982

[11] Azamjon Muminov, Yun Chan Jeon, Daeyoung

Na, Cheolwon Lee, Heung Seok Jeon,

“Development of a Solar Powered Bird

Repeller System with Effective Bird Scarer

Sounds,”, ICISCT., pp.85-87, 2017,

doi:10.1109/ICISCT.2017.8188587

[12] R. YAURI et al., “Development of an

Electronic Bird Repellent System using Sound

Emission” WSEAS TRANSACTIONS on

SYSTEMS and CONTROL, vol. 18, 2023, doi:

10.37394/23203.2022.18.14.

[13] Fuliang Le, Jiawei Luo, Gongping Wu, ”An

Uninterrupted Bird Repeller on Transmission

Line,” International Conference on Robotics

and Biomimetics, pp.19-23, December, 2009

[14] EZEONU Stella Ogochukwu, AMAEFULE

Desmond Okechukwu, OKONKWO Godfrey

Nnaegbo, “Construction And Testing Of

Ultrasonic Bird Repeller,” Journal of Natural

Sciences Research Vol.2, No.9. pp.8-17, 2012

[15] Siyang Li, Xingguang Li, Zhaoliang Xing,

Zhuo Zhang, Yifan Wang, Ruixiao Li, Runting

Guo, Jiangjian Xie, “Intelligent audio bird

repeller for transmission line tower based on

bird species variation,” IOP Conf. Series:

Materials Science and Engineering, pp.1-

4,2019, doi:10.1088/1757-899X/592/1/012142

[16] Armin R. W. Elbers, José L. Gonzales.,

“Efficacy of an automated laser for reducing

wild bird visits to the free range area of a

poultry farm,” Scientifc Reports, 2021,

Doi:10.1038/s41598-021-92267-z

[17] Kohei Maruyama, Takeshi Saitoh and Ryosuke

Konishi, ”Bird Tracking and Flapping Motion

Recognition for Monitoring System,” ICROS-

SICE International Joint Conference, pp.18-21,

August 2009, Fukuoka International Congress

Center, Japan

[18] Lucas A. Bluff and Christian Rutz, ”A quick

guide to video-tracking birds,” Biol. Lett. vol.

4, pp.319-322,2008

[19] Samuel Matos, Raul Morais, P.M. Araujo,

P.J.Q. Tenreiro, P.J.S.G. Ferreira, M. J. C. S.

Reis, “A GSM-based System for the Tracking

of Birds, ” SENSORDEVICES 2015 :The Sixth

International Conference on Sensor Device

Technologies and Applications, pp.131-137,

2015

[20] R.Vishnuvardhan, G.Deenadayalan,

M.V.Vijaya Gopala Rao, Sheetal P Jadhav,

Adithya Balachandran, ”Automatic detection of

flying bird species using computer vision

techniques,” International Conference on

Physics and Photonics Processes in Nano

Sciences, Journal of Physics: Conference

Series, pp.1-7, 2019, Doi:10.1088/1742-

6596/1362/1/012112

[21] Santosh Bhusal, Shivam Goel, Kapli Khanal,

Matthew Taylor, Manoj Karkee, “Bird

Detection, Tracking and Counting in Wine

Grapes,” An ASABE Meeting Presentation,

pp.16-19, Jul. 2019.

[22] Dulari Bhatt, Chirag Patel, Prof. Priyanka

Sharma, “INTELLIGENT FARM

SURVEILLANCE SYSTEM FOR BIRD

DETECTION,” G.J. E.D.T., vol. 1, no.2, pp.

14-18, 2012

Contribution of Individual Authors to the

Creation of a Scientific Article (Ghostwriting

Policy)

All authors have contributed equally to the creation

on this paper.

Sources of Funding for Research Presented in a

Scientific Article or Scientific Article Itself

No funding was received for conducting this study.

Conflict of Interest

The authors have no conflict of interest to declare.

EARTH SCIENCES AND HUMAN CONSTRUCTIONS

DOI: 10.37394/232024.2024.4.3

Subong Pyon, Jongsu So, Junhyok Pak,

Jinzu So, Sunil Kim

E-ISSN: 2944-9006

Volume 4, 2024

Creative Commons Attribution License 4.0

(Attribution 4.0 International, CC BY 4.0)

This article is published under the terms of the

Creative Commons Attribution License 4.0

https://creativecommons.org/licenses/by/4.0/deed.en

_US

EARTH SCIENCES AND HUMAN CONSTRUCTIONS

DOI: 10.37394/232024.2024.4.3

Subong Pyon, Jongsu So, Junhyok Pak,

Jinzu So, Sunil Kim

E-ISSN: 2944-9006

Volume 4, 2024