Tóm tắt luận án tiếng anh: Nghiên cứu công nghệ quét laser mặt đất trong lĩnh vực địa hình và phi địa hình ở Việt Nam.

MINISTRY OF EDUCATION AND TRAINING
HANOI UNIVERSITY OF MINING AND GEOLOGY

CUONG SY NGO

RESEARCH ON TERRESTRIAL LASER
SCANNING TECHNOLOGY IN TERRAIN AND
NON-TERRAIN FIELDS IN VIETNAM

Major: GEOMATICS AND MAPPING ENGINEERING
Code: 9.52.05.03

SUMMARY OF THE PHD THESIS

HANOI, 2022


This thesis was completed at Department of Photogrammetry and
Remote Sensing, Faculty of Geomatics and Land Administration,
Hanoi University of Mining and Geology

Supervisor
Assoc. Prof. Dr. Truong Xuan Tran,
Hanoi University of Mining and Geology.

Reviewer 1: Prof. Dr. Quy Ngoc Bui
Reviewer 2: Prof. Dr. Hung Le Trinh
Reviewer 3: Dr. Khang Du Nguyen

The thesis will be defended at the University-level thesis evaluation
committee at Hanoi University of Mining and Geology, at 8.30 am

on.......................2022

The thesis can be found at theses libraries: The Hanoi
National Library or The library of Hanoi University of
Mining and Geology


1
INTRODUCTION
1. Rationale of the study
Nowadays, in the field of ground space data collection, the
integration of Global Positioning System (GPS), digital image
technology and laser technology for data exchanges via the Internet
is a trend. The typical one is the laser scanning system. The
scientific, complete and systematic research in order to bring 3D
Terrestrial Laser Scanning (TLS) technology in the fields of terrain
and non-terrain is the interest of scientists, researchers and managers.
Therefore, the PhD thesis’s topic ‘Research on Terrestrial Laser
Scanning technology in terrain and non-terrain fields in Vietnam’
was selected based on the scientific and practical significance.
2. Research aims and tasks

2.1. Aims: To establish a scientific basis and methodology for
building ancillary programs in Terrestrial Laser Scanning data
processing. To building 3D TLS technological processes for several
non-terrain objects.
2.2. Tasks: Research on a theoretical basis of 3D TLS technology
and error sources affecting the accuracy; research the algorithms and
develop a program to process the 3D TLS data; experiment to verify
the research results.

3. Objects and scope of the study
3.1. Objects of the study: The Digital Topographic Models (DTM) of
the areas with diverse topography was established, and the accuracy
was assessed. In terms of the non-terrain field, the 3D models of
caves, landscape streets and historical and cultural relics were built,
then modeling for the socioeconomic development.
3.2. Scope of the study: The terrain objects are the areas of Gia Phu
village in Binh Duong commune, Gia Binh district; and the Lim hill
of Lim town in Tien Du district, Bac Ninh province. Non-terrain
objects are the Dau Go cave in Ha Long Bay, Quang Ninh province;


the Ta Hien and Dinh Liet old streets in Hoan Kiem district, Hanoi
capital; the Buu Long pagoda in Ho Chi Minh city; and some
antiques of the But Thap pagoda in Bac Ninh province.
4. Research methods
Statistical method; analysis and synthesis methods; modeling
method; and experimental method.
5. Scientific and practical significance of the thesis

5.1. Scientific significance: Building and completing in a
synchronous and scientific way the theoretical basis of 3D TLS
technology. Proposing the solutions, technological processes, and the
auxiliary programs for the application of 3D TLS in the geospatial
data collection and processing. The purpose then is for building the
DTM and a number of non-terrain applications in the actual
conditions of our country.
5.2. Practical significance:
The 3D TLS was brought into play the advantages in maximize
automation of data collection and processing to meet the

technical, quality and cost requirements. The point filtering
program was introduced and successfully applied in processing
the 3D TLS data in accordance with practical conditions in our
country. Then, these results will contribute to promote the
applications of new technologies and to serve the goal of
modernizing the cartography field.
6. Defense points
Defense point 1: Using the scanning data filtering algorithm with the
addition of surface modeling, developing a TIN model and
enhancing the filters in order to build an ancillary program. This one
allows automating point cloud data processing, ensuring the quality
as well as the efficiency in the DTM establishment.
Defense point 2: Proposing the 3D TLS technological process for
building 3D models for different non-terrain objects synchronously.


It ensures the technical requirements of the accuracy to meet the
socioeconomic development requirements in Vietnam.
7. New points of the thesis
- Building the ancillary program in 3D TLS data processing based on
the scanning data filtering algorithm with the addition of surface
modeling, developing a TIN model and enhancing the filters. It will
help to accurately represent the DTM accordance with current
requirements in Vietnam.
- Proposing the 3D TLS technological process for building 3D
models for different non-terrain objects.
8. Thesis structure and content
Introduction: An overview of the thesis, the rationale, aims,
significance, and literature review in local country and in the world
related to the content of the thesis. From there, the research methods

and the content are formed. At the same time, the defense points and
the new points of the thesis are given.
The main research contents of the thesis concludes four
chapters:
Chapter 1: Overview of literature review of 3D Terrestrial Laser
Scanning Technology in the terrain and non-terrain fields.
Chapter 2: Scientific basis and error sources affecting the results of
Terrestrial Laser Scanning.
Chapter 3: The program establishment to process the 3D Terrestrial
Laser Scanning data in the terrain field.
Chapter 4: Process establishment of 3D Terrestrial Laser Scanning
for surveying several non-terrain objects.
Conclusion and recommendation part: Summarizing the research
parts in the thesis and giving the conclusions affirming the research
results on the point cloud data processing filtering by TLS.
Proposing the 3D TLS technological process for surveying each
specific non-terrain object.


Chapter 1
OVERVIEW OF LITERATURE REVIEW OF 3D
TERRESTRIAL LASER SCANNING TECHNOLOGY IN THE
TERRAIN AND NON-TERRAIN FIELDS
Within the scope of this thesis, the author only focuses on research
of the 3D TLS technology. Laser technology was conceived and
developed in the 1960s by the scientist namely Theodore H. Maiman.
It is based on the basis of the theory of Charles Hard Towner and
Asthur Leonard Schawlow [76].
Literature review of the 3D TLS studies
1. Literature review in the world

The application of the 3D TLS technology in terrain and non-terrain
fields has been concerned by many researchers around the world,
mainly in the field of non-terrain applications.
1. Research and applications in terrain field: There are a number of
researchers such as Nora Tilly, Peter Schürch, Andri Baltensweiler,
etc. [30,31].
2. Research and applications in non-terrain field: In the research of
Glenn J. Newnham, Francesco Pirotti, Mathieu Dassot, Mohammed
Oludare
Idrees,
Soohee
Han
and
Yusuf
Arayici
[12,13,14,27,35,36,40].
3. Research and application of point filtering algorithms: Scientists
have done including L.Monika Moskal and Guang Zheng [18],
J.Susaki [39].
4. Research and commercial software development: such as
TRIMBLE [72], LEICA [73], RIEGL [74], FARO [75].
In addition, the problems of the TLS accuracy is also concerned by
many scientists around the world. For example, author group J.Angelo Beraldin, Franỗois Blais, Uwe Lohr [16], or group of authors
Balzani, M., Pellegrinelli, A., Perfetti, N., Uccelli, F. [10].


In general, the problems of current TLS technology are quite complete.
However, in these studies, the authors only focused on scientific
information. They did not give any specific technological processes and
did not analyze the detailed algorithms. These references are only basic

information, mainly focusing on specific applications, and often quite
fragmented, difficult to absorb and implement.
2. Literature review in Vietnam
The applications of 3D TLS in the terrain and non-terrain fields are
quite cared by scientists in our country. In 2020, the author Nguyen
Viet Nghia established a coal mine map with the high accuracy using
the TLS technology. In 2019, the Ministry of Natural Resources and
Environment issued Circular 17/2019/TT-BTNMT namely
"Technical regulations for establishing DEM by ground-based
LiDAR scanning technology on fixed stations".
In 2017, the author Tran Quoc Vinh researched combining data of
drones and TLS in order to create the 3D maps of urban areas [7]. In
2015, the author Vu Quoc Lap chaired the topic namely "Research
and application of TLS technology to improve the geospatial data
quality , then to strengthen the state management capacity in mineral
activities" [4]. The application of TLS technology to terrain and nonterrain purposes is not new in our country, but it is still not
systematic.
Sub-conclusions of Chapter 1
In this chapter, the author has stated the reasons for choosing the 3D
TLS technology for terrain and non-terrain applications. This is the
new technology in Vietnam following the trend of 4.0 technology
revolution.
This chapter was presented some applications of the 3D TLS
technology in terrain and non-terrain fields. It was also presented
the literature review using TLS technology in the world and in
Vietnam. There are only some research and applications in the field


of terrain in our country currently, and the applications in the field
of non-terrain is still limited. From there, the author draws out the

remaining problems that will be solved in the thesis.
Chapter 2
SCIENTIFIC BASIS AND ERROR SOURCES AFFECTING
THE RESULTS OF TERRESTRIAL LASER SCANNING

2.1. Working principles of terrestrial laser scanning system

2.1.1.

General principles
Figure 2.2. Functions and main components of Terrestrial Laser
Scanning equipment [62]

1 - long measuring block; 2 - laser rotary optomechanical device block; 3 transmission line - long measurement signal receiver; 4 - channel of data
transmission to the computer; 5 - dedicated computer.

Figure 2.3. Coordinate system of TLS stations [20]


2.1.2. Working principle of the long measuring block of the TLS
equipment

1. Determination of distance length by pulse measurement “Time of flight”
Figure 2.4. Principle of pulse measurement to determine distance
length [59]

2. Determination of distance length by phase measurement “Phase based”
Figure 2.5. Diagram to determine distance length by
phase measurement [55]
3. Determination of distance by laser scanning triangulation method

Figure 2.6. Distance determination by laser


scanning triangulation method [52]


2.1.3.

Working principle of TLS equipment rotary block

1. Working mode of step rotation of TLS equipment
Figure 2.7. Principle of angle positioning step [52]

2. Working mode of continuous rotation of TLS equipment
Figure 2.8. Principle of the continuous angle positioning [52]
2.2. Error sources affecting TLS results
2.2.1.
Classification of TLS errors
All errors in TLS can be divided into two groups:
1. Equipment errors are caused by the quality of manufacturing,
assembling and adjusting optical, mechanical and electronic parts of
scanning equipment. They include as follows: working stability of
TLS equipment; working accuracy of long measuring block; errors
of angle blocks. These errors are shown in the profile of the scanning
device during the manufacturing, assembly process, and are
periodically checked.
2. Method errors that the measurands themselves are determined by
TLS equipment. Methodological errors caused by the surrounding

environment when doing the laser scanning; Methodological errors
are caused by the characteristics of the scanned objects. Moreover,
there are also errors during the coordinate transformation, errors in
data conversion.
2.2.2.
Methodological errors of TLS
1. Surrounding environmental effects when TLS:
The influence of factors includes atmosphere, refraction, attenuating
electromagnetic waves, device vibration, etc. To determine the
refractive index of electromagnetic waves taking into account
meteorological data when scanning objects, the formula [61] can be
used:
��−1
�
55�
x 10−9
(2.19)
�� = 1 +1+ (�+273,16) 760-1+
(�+273,16)
where, α - the gas constant và equal to 1/273.16;
P - the atmospheric pressure, measured in mm of mercury
column;
E - the absolute humidity of the air, unit of measure
3
gram/m ;
T - the air temperature, unit of measure Kelvin;
nG - the atmospheric standard refractive index calculated
according to the formula [61]:
16,288

7
(�
+
(2.20)
� – 1) . 10 = 2876, 04 +
0,136
�4
�
λ – the radiation
2
wavelength.
2. Errors caused by the characteristics of the scanned objects:
The influence on the TLS results shows that the basic properties of
the scanning objects are shape and reflectivity, which are determined
by the material composition and color. The influence of the
measuring objects on the accuracy of received spatial information,
which mainly appears about the errors of the measuring distances.
The process of transmitting laser pulses into space is very
complicated.


The input signal function has the following general form:
� (t)
∫ ∆ [ (� − 2�(��, ) ) � (�,� ) �(�, �)]
�
��
�

−∆

−∆

�
�

�

d�. D� +

���

η(t)
(2.23)
where, Fout (t) – the output signal function;
D (φ, υ) – the function describing the change from the laser
emitted source to different points of the object (reflecting the shape
of the object);
Kref (φ, υ) – the function with reflective properties that
calculates the signal reflectance from different points of the object;
ψ (φ, υ) - the function describing the signal distribution in
space, the laser dispersion quantity;
ɳ(t) - the function with the noise component characteristic of
the scan results.
3. Influence of scanning parameters and scanner characteristics on
the accuracy of measurement results
When doing the TLS, it is necessary to determine the characteristics
of the TLS stations, including: maximum vertical and horizontal
scanning resolutions; laser radiation wave length; edge, vertical and
horizontal measurement accuracy; dispersion of the laser beam.
4. Errors in data reference conversion

The errors are related to the data acquisition and feedback of scan
rays from objects to refer to the geodetic reference system. These
influencing factors will be described in details on data quality [25].
5. Errors in data format conversion
Point cloud data usually has standard formats such as: e57, ptx, pts,
fls, lfc, etc. Each scanner manufacturer creates its own standard
format. If the third-party users want to use the data, they must
perform a conversion process. This is also an important point during
the data processing and exchange.
Sub-conclusions of Chapter 2


TLS technology operates on the foundation of science and
technology and information technology development. It affirms its
superiority compared with other traditional technologies. It
contributes to strengthening and modernizing the surveying,
mapping and geographic information fields. The research content of
this chapter contributes to supplementing and perfecting a number of
issues on the theoretical basis of TLS for teaching, scientific
research as well as pratical applications in our country.
Through analyzing and evaluating the error sources affecting the
TLS results, it is found that the errors in the coordinate conversion
process has a large and systematic influence on the accuracy of the
point cloud. Then, it is considered as a basis for providing solutions
to handle requirements of collected quality data and to choose the
suitable equipment when performing different work purposes.
The data obtained from the TLS is large and diverse. There is a need
to build an auxiliary program allowing the automation of point cloud
data processing and the model quality assurance for terrain and
non-terrain aims. This issue is addressed in the next chapter.

Chapter 3
THE PROGRAM ESTABLISHMENT TO PROCESS
THE 3D TERRESTRIAL LASER SCANNING DATA
IN THE TERRAIN FIELD
3.1. Filtering algorithm from point cloud data to build DEM
To classify land surface and non-land surface layers, there are many
filtering algorithms to optimize automation in the point filtering
process. This algorithm is grouped according to 3 basic criterias
which are [35]:
 Slope - topography




Geo morphological model
Land cover




X

Step 2:
Parameter
estimatio

Step 3:
Filtering
DTM original
data

Preparing DTM input data

Parameter estimation using statistic analysis
The DTM results

Figure 3.1. Flow chart of the surfaceEstimated
modeling
filter algorithm,
parameters
the development of TIN model and filter enhancement [28,42]
Developing TIN models and enhancing the filters


The process of filtering the DTM data on the basis of developing the
TIN models and enhancing the filter is carried out in 6 steps:
Finding similarity (same name) of TIN model based on
Step 1:
the appearance of unfiltered points.
Step 2:

Calculation of the slope, topographic objects of TIN
model.
Manipulating the symmetric points, vertices, corners,
filtered data with labels Pvertex (Xv, Yv, Zv), and with the
highest elevation value of the TIN model and unfiltered
points namely Punfiltered (Xu, Yu, Zu,).

Step 3:

The symmetries (similarities) are P mirror (Xm, Ym, Zm) of
Punfiltered (Xu, Yu, Zu,) are calculated by the formula:
�� = 2�� − ��
{ �� = 2�� − ��
(3.12)
�� = � �
Calculating the d distance and the θ angle of the TIN
model with the points filtered to Vertex form. If both

Step 4:

calculated values are less than the maximum values d
and θ, then the unfiltered point will be the label
(Vertex) of the ground point.
Calculating the scale of the TIN model on the plane
with coordinates X and Y. If the coefficient (RATIO) of

Step 5:

the max - min scale is smaller than the scale of the
similarities, then it will be updated additionally to the
DTM input data.

Step 6:

Repeating the step 1 to step 5 until all the unfiltered
points are are added to the final DTM result.

3.2. Building an automatic point cloud filter processing program

3.2.1.
Flow chart of the program
(Figure 3.2. Flow chart of the point cloud data processing program)
3.2.2.
1.
2.
3.
4.

Steps of the flow chart
Preparing and normalizing the input data
Reducing the point cloud density
Filtering manually the data
Filtering the terrain data

5. Exporting data to LAS file
Figure 3.3 Flow chart of the supporting program for point cloud
data processing


3.3. Experiment to compare and assess the accuracy of DEM
after using the automatic point cloud filtering
3.3.1.
The case study
The case studies are the Gia Phu village of Binh Duong commune, Gia
Binh district, and the Lim hill of Lim town, Tien Du district, Bac Ninh
province. In addition, about 10 feature points in each area were selected
and the coordinates and the altitudes were determined accurately by
total stations. They are a basis for comparing and evaluating the

accuracy with the data after processing the above program.
3.3.2.
Experimental results and accuracy assessment
1. The Gia Phu village area

Figure 3.4. Point cloud model
after rectifying

Figure 3.5. Point cloud after
color assignment of the Gia Phu
village area

Figure 3.6. Remaining terrain
data after filtering

Figure 3.7. Non-terrain classified
from the original data


Figure 3.8. Point cloud collected
by UAV method

Figure 3.9. Available terrain
data area

Comparing the accuracy between two data types:
Table 3.1. Comparison table of TLS data and available terrain
data

Accuracy assessment with the field points

Table.2. Evaluation table of absolute errors of the feature objects
Altitude Data
Direct
Location ID
X
Y
obtained commeasurement
from parison
(m)
(m)
altitude (m)
TLS (m) (m)
TD1
578170.868 2330807.078
2.893
2.859 -0.034
TD2
578150.874 2330803.373
2.811
2.897 0.086
TD3
578175.457 2330812.192
2.834
2.828 -0.006
TD4
578182.261 2330816.364
2.600
2.573 -0.027
TD5
578187.926 2330829.649

2.867
2.854 -0.013
TD6
578168.703
2330818.6
2.907
2.813 -0.094
TD7
578179.544 2330803.926
2.684
2.616 -0.068
TD8
578166.161 2330804.428
2.949
2.926 -0.023
TD9
578170.825 2330816.719
2.887
2.841 -0.046
TD10
578187.432 2330819.292
2.963
2.846 -0.117


* Altitude mean square errors:
Accuracy assessment between TLS and UAV methods:
∑(∆ℎ)2 = ± 0.033 (m) (3.14)
= √ �
Accuracy assessment between TLS and direct measurement methods:

�ℎ

∑(∆ℎ)2 = ± 0.063 (m)
= √ �
From the results of the mean square errors (Formula (3.14)) of the
feature points, it shows that the altitude mean error �ℎ <1/4 of the

�ℎ

vertical interval. That proves that the point cloud filtering program
when using TLS ensures the accuracy of the 1/500 scale DEM with a
vertical interval h = 0.5m.

2. Lim hill area

Figure 3.10. Point
cloud collected by
TLS in the Lim hill
area

Figure 3.11. DEM
of the case study

Figure 3.12.
Terrain data in
Lim hill by
Lidar

Table 3.3. Elevation comparison table of two methods of DEM
building

No

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

26

Altitude
obtained
from
Lidar
(m)
22.255
22.363
21.702
21.554
21.749
21.825
16.009
22.406
22.417
21.145
16.976
22.433
17.578
21.843
21.479
20.986
18.733
15.359
19.01
18.278
18.794
15.385
16.323

18.896
15.081
15.19

Altitude
Altitude
Altitude
obtained
obtained Deviant
obtained Deviant (m)
No from
from
(m)
from
Lidar
TLS (m)
TLS (m)
(m)
22.722 0.467
-0.025
27 13.696 13.671
22.119 -0.244 28 14.647 14.872
0.225
22.024 0.322
29 13.656 13.64
-0.016
21.798 0.244
0.236
30 13.176 13.412
21.605 -0.144 31 13.085 13.19

0.105
21.963 0.138
32 11.885 11.651
-0.234
16.348 0.339
-0.216
33 13.37 13.154
22.731 0.325
34 11.656 11.845
0.189
22.051 -0.366 35 11.547 11.631
0.084
21.279 0.134
36
11.5
11.483
-0.017
16.524 -0.452 37 9.555
9.753
0.198
22.176 -0.257 38 11.315 11.452
0.137
17.83
0.252
-0.247
39 10.426 10.179
21.73
-0.113 40 9.771
9.637
-0.134

21.661 0.182
41 9.752
9.582
-0.170
20.678 -0.308 42 9.326
9.477
0.151
18.479 -0.254 43 9.621
9.586
-0.035
15.656 0.297
9.577
-0.214
44 9.791
19.369 0.359
0.191
45 10.101 10.292
18.48
0.202
46 9.458
9.545
0.087
18.402 -0.392 47 9.384
9.529
0.145
15.536 0.151
9.53
0.092
48 9.438
15.991 -0.332 49 10.167 10.285

0.118
18.429 -0.467 50 9.221
9.508
0.287
15.243 0.162
51 9.239
9.501
0.262
15.027 -0.163


* Altitude mean square errors:
∑(∆ℎ)2 = ± 0.240 (m)
(3.15)
= √ �
where, ∆ℎ is the altitude difference; n is the total number of tests (n
= 51). From the results of the mean square errors according to the
formula (3.15) of the characteristic objects, it shows that the altitude
mean error �ℎ <1/4 of the vertical interval. That proves that the
point cloud filtering program when using TLS ensures the accuracy
of the 1/2000 scale DEM with a vertical interval h = 1 m.
Sub-conclusions of Chapter 3
The study of point filtering algorithms, then to propose automatic
and semi-automatic algorithms combined with the selection of
suitable parameters to create DEM from point cloud data, is topical,
updating, modern, and highly practical. The main result of this
chapter is to provide a program to support point cloud data
processing. The experiment was confirmed to meet the requirements.
The program is Vietnameseized, easy to use and suitable for mass
production. Thereby, it will help to reduce and shorten the time to do

the office work and to save the costs.

�ℎ


Chapter 4
PROCESS ESTABLISHMENT OF 3D TERRESTRIAL LASER
SCANNING FOR SURVEYING SEVERAL NON-TERRAIN
OBJECTS
4.1. Technological process to build 3D models of non-terrain 3D
objects by TLS

Figure 4.1. Process of surveying non-terrain objects


4.2. Experiment of 3D TLS of cave objects
It is Dau Go cave on Dau Go island in Ha Long Bay, Ha Long City,
Quang Ninh province.

Figure 4.2. Scanning station and target arrangement in the field at Dau
Go cave

Figure 4.3. The 3600 image
extracted from a TLS station

Figure 4.4. Experimental
area point cloud data

Figure 4.5. Point cloud model of
Dau Go cave area

Figure 4.6. Terrain model of the
case study


4.3. Experiment of 3D TLS of old street objects
Collecting data by TLS for including Ta Hien and Dinh Liet streets,
Hoan Kiem district, Hanoi City.

Figure 4.7. Point cloud model
of Ta Hien Street

Figure 4.8. Point cloud model
of Dinh Liet Street

Figure 4.9. The 3600 picture of Ta Figure 4.10. Report of Ta Hien
Hien street
experimental street
4.4. Experiment of 3D TLS of cultural heritage objects

Figure 4.11. Experimental map of
the Buu Long relic pagoda

Figure 4.12. The model of
Buddha with a thousand
eyes and hands– But
Thap
pagoda