Integrating UAV Photogrammetry and Terrestrial Laser Scanning for
the 3D surveying of the Fortress of Bashtova
Abstract: - Through the synergistic application of Aerial Photogrammetry using UAVs and Terrestrial Laser
Scanning (TLS), this paper investigates how this combination can be used for conducting a 3D survey of the
Fortress of Bashtova thereby demonstrating the effectiveness of such integrated methods in acquiring an all-
encompassing image of this historical building. As the efforts towards preservation become intense, there arises
the urgency of precise and detailed 3D documentation that will facilitate appropriate conservation processes
and further studies. Therefore, combining TLS and UAV photogrammetry offers a powerful tool that can
provide accurate architectural data for the documentation of heritage areas. Moreover, the TLS component
acquires ground point-cloud data with laser scanners giving a complementary alternative for aerial perspective.
The merging of these datasets ensures broad inclusion since it allows the production of accurate, detailed three-
dimensional models of the Fortress of Bashtova. Thanks to the research on the case study of the Fortress of
Bashtova in the article, it can be stated that the integration of data from aerial photogrammetry and TLS is
seamless with the help of modern software while respecting the basic photogrammetric-geodetic rules and
demonstrates the possibility of creating a complex 3D model, usable for further analyses for architects and
conservation professionals, as well as for restorers and civil engineers. To estimate the accuracy of the point
clouds derived from TLS and UAV, we compared the distances between the point clouds using CloudCompare
software. We obtained a mean RMS of 2.199073 mm and std. dev was 7.356 mm. Research has shown that the
difference between point clouds from TLS and UAV is within 1.7 centimeters.
Key-Words: - UAV, Terrestrial Laser Scanning, 3D Model, Heritage, Photogrammetry, Point Cloud, Accuracy.
Received: July 24, 2023. Revised: April 24, 2024. Accepted: June 11, 2024. Published: July 10, 2024.
1 Introduction
Today, the widely accepted technologies must be in
a place to uncover such historic buildings and how
such mysteries have been solved as archeological
documentation and conserving of historical remains
cannot do without it. Among these, the combination
of terrestrial laser scanning (TLS) and aerial
photogrammetry stands out as a revolutionary
method that provides a cooperative response to the
difficulties presented by the elaborate architecture
and vast landscapes of castles. Combining
Terrestrial Laser Scanning (TLS) and Aerial
Photogrammetry within this context emerges as one
forerunning method capable of providing a
complementary answer to such challenges brought
about by the detailed structures and large
environments found at castle sites.
Monuments and other items that are part of
cultural heritage have long been documented using
geodetic methods. The speed and precision of
documentation work have greatly risen with the
advancement of computer technology and new
equipment. Photogrammetry was invented and
photography started to be used for documenting
about 150 years ago. Following World War II,
electronic methods were progressively included in
surveying, and in the 1970s, satellite data started to
be employed in addition to aerial images. With the
advent of widely accessible computer technology
and the digitization of technology, the 1990s saw a
significant shift. These days, automated close-range
photogrammetry from the ground and drones,
airborne systems, satellite systems, terrestrial and
mobile laser scanning, and electronic surveying
systems (total stations, GNSS equipment) are used.
Research on the synergy of data from many
instruments is being conducted at many workplaces
these days.
In this paper, we will investigate the use of this
integration technique using Bashtova Castle as our
main example. The rich historical significance and
the complex architectural features of the Bashtova
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.30
Arli Llabani, Otjela Lubonja
E-ISSN: 2224-3496
306
Volume 20, 2024
1ARLI LLABANI, 2OTJELA LUBONJA
1Faculty of Civil Engineering, Polytechnic University of Tirana, Rruga Muhamet Gjollesha,
ALBANIA
2Faculty of Engineering, Informatics and Architecture, European University of Tirana, Rruga Xhanfize Keko,
ALBANIA
Fortress provide an interesting platform from which
to test if it is possible to combine Aerial
Photogrammetry and TLS for full 3D surveying of
castles.
Castles are intriguing not just because of their
past, but also because they are constructed with lots
of fine points that residents and visitors find equally
fascinating usual surveying procedures rarely take
into consideration all aspects when it comes to
castles. Nevertheless, aerial photogrammetry
utilizing unmanned aerial vehicles (UAVs) together
with high-resolution images constitutes the best
method through which one can get an overview of a
castle from above, [1].
At the same time, we have the terrestrial laser
scanner that is based on the ground targeting
detailed printouts of the inside parts of the fortress
together with its structural characteristics. The
combination of these methods can bring a detailed
3D survey of cultural heritage areas including all
details and eliminating information gaps, [2], [3].
This study presents the integration of UAV
photogrammetry and terrestrial laser scanning for
the 3D survey of the Fortress of Bashtova, as well as
presents a methodology for the management of
cultural and archaeological areas, [4].
The results of this scientific research show that
the integration of these two methods for the 3D
survey of cultural heritage areas provides an
effective approach to capturing all the architectural
elements and details with high accuracy and
provides a comparison of the accuracy between
UAV photogrammetry and terrestrial laser scanning,
[5], [6].
2 Materials and Methods
2.1 Case Study
We used the Fortress of Bashtova as a case study in
this paper. Bashtova castle is located at a distance of
3 - 4 kilometers near the village of Vile-Bashtove, in
the north of Shkumbini river.
This castle was built in the 15th century and is a
beautiful testimony to the civilizations that have
passed through Albania. It is 36 kilometers north of
Fier, 20 kilometers northwest of Lushnja, 15
kilometers south of Kavaja, and 40 kilometers
southwest of Tirana as shown in Figure 1.
It is the only castle in the Balkans to be
constructed on a field.
Fig. 1: Location of the Fortress of Bashtova
2.2 Architectural Analysis
The Fortress of Bashtova has a rectangular plan
measuring 60 x 90 m. In the four corners and the
middle of each wall, there is a tower, except for the
western wall which belongs to a second construction
period. The walls have a width of 1m. Between the
sandstone and conglomerate in irregular shapes,
pieces of bricks and tiles have been inserted here
and there. From the inside, the walls are broken by a
system of pilasters with a section of about 1x 1.5m
at every 3m distance.
In the upper part, the pilasters relate to brick
arches, 0.40 m high, strengthened with wooden ties,
and create the arches over the guard path 1.20 m
wide. The walls' total height, including all of the
beams, was nine meters. The height of the
balustraded parapet is 1.90 meters, and its width is
0.80 meters. Arrows measuring 0.50 meters in
height, 0.35 meters in internal width, and 0.15
meters in external width are used to characterize
each bar.
The niches, which are created between the
pilasters, are also equipped with two rows of friezes.
The towers have circular or quadrangular shapes.
Two of the corner towers are round, and one is
quadrangular, while no traces of the fourth are
preserved as shown in Figure 2.
The intermediate towers are all quadrangular.
They have a wall thickness of 1.25-1.40 m, while
their height reached 12 m. These premises were not
inhabited but served only in times of war. In a later
period, some tower areas were used for living, being
equipped with fireplaces.
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.30
Arli Llabani, Otjela Lubonja
E-ISSN: 2224-3496
307
Volume 20, 2024
Fig. 2: The plan of the Bashtova Castle
The castle had three entrances, as can be judged
from the preserved traces. The gate was covered
with architraves and had a light space of 2.70 m,
while in a later period, this gate was closed with a
wall. The stairs are built of stone walls and rest on
the inside of the walls, outside their thickness. The
architecture and construction technique point to a
fort built in haste, considering the greatest saving of
material.
This is evidenced by the thin walls, the harbor
combined from the inside with pilasters and arches,
the open towers that are less resistant as well as the
low floors of the towers designated only in case of
wars.
Over the years, several conservation
interventions have been undertaken in the Bashtova
Castle, relying on the materials found in the Cultural
Monuments archive.
2.3 UAV Photogrammetry
Drones are unmanned aerial vehicles (UAVs) that
operate without a pilot. Tһe mapρing, mοnitoring,
аnd mіlitary ѕectorѕ have sееn wideѕpгeаd
utilization of unmanned aегial vehicles (UAVs) due
to their аbility to click high-resolution data
stagecoach versatility. In theory, the UAV system
comes with sensors (such as a camera, LiDAR, or
thermal sensor), navigational aids, and
communication tools, [7], [8].
A DJI Matrice 300 RTK with a Zenmuse P1
camera was utilized during an aerial review of
Bashtova Castle in this probe. The DJI Matrice 300
RTK is powered by a quadcopter propulsion system
with four rotors for lifting and propulsion. These
motors are high-performance, brushless, and are
designed to provide the best power and efficiency
whilst remaining reliable and silent. Additionally,
the DJI Matrice 300 RTK has an integrated obstacle
avoidance technology that makes use of advanced
sensors and algorithms to detect and avoid things
blocking its way, while at the same time being
equipped with a strong GPS for accurate positioning
and steering, [9].
We used 5 GCP with a width of 60 cm and a
length of 60 cm as shown in Figure 3. These points
served us to increase the accuracy of the images
taken during the photography of the castle with both
methods. Ground Control Points (GCPs) are large,
easily recognized photo targets that are positioned
on the ground inside the drone survey's perimeter.
Their purpose is to make sure that each point's
coordinates in the images most closely match the
GPS coordinates with high accuracy, [10].
For this study, we marked 5 GCPs and
measured them with a Trimble R12i receiver,
obtaining RTK data from the Albanian National
GNSS System "ALBCORS.
Ground control points must always be visible
during aerial photography and this is achieved by
using high-contrast colors and making sure the size
of the control points is visible enough for the flight
altitude we are working at.
Fig. 3: Ground Control Point distributed around the
castle
Ground Control Point coordinates were acquired
in the UTM Zone 34N coordinate system
(epsg:32634) with an absolute precision of 1 mm as
shown in Table 1.
Table 1. Coordinates of 5 Ground Control Points
measured with Total Station
GCP
X (m)
Y (m)
H (m)
1
4545029.727
373683.198
3.474
2
4545085.396
373671.341
3.246
3
4545093.246
373599.915
3.224
4
4545012.658
373619.673
3.012
5
4545055.091
373644.263
3.491
We used DJI Matrice 300 RTK with Zenmuse
P1 which has the largest image sensor with the
highest resolution ever. The Zenmuse P1 camera
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.30
Arli Llabani, Otjela Lubonja
E-ISSN: 2224-3496
308
Volume 20, 2024
also supports prime lens swapping. These lenses are
capable of producing 45-megapixel aerial photos,
and the camera is mounted on a 3-axis gimbal. The
DJI Zenmuse P1 with a 35mm focus lens was tested
for this study.
For UAV Photogrammetry was performed
oblique mission using DJI Pilot 2 which involves a
main flight path to collect nadir photos in addition
to multiple subpaths facing towards the center of the
site to collect oblique photos. This method requires
more flight time and battery power than a standard
2D Area Route mission of the same site, [11], [12].
For the oblique area route, 5 different missions
were performed to capture all the details of the
castle, with different camera angle positions. All 5
missions, then were merged into a project, to
generate a Point Cloud with high accuracy, [13],
[14], [15].
We choose 50 m height and 3 m/s speed of DJI
Matrice 300 RTK to perform this flight path as
shown in Figure 4. A front overlap of 85% and a
side overlap of 80% were set.
Fig. 4: Oblique mission performed with DJI Matrice
300 RTK
2.4 Terrestrial Laser Scanning
Terrestrial Laser Scanning, otherwise described as
TLS, is a version of laser scanning that uses Light
Detection and Ranging for the creation of a 3D
point cloud. In a brief amount of time, it can gather
millions of points. Because of this, it has been
applied to a wide range of tasks, including part-built
and as-built model creation, progress control,
change detection, building diagnostics, and project
monitoring, [16], [17].
Static TLS and mobile laser scanning (MLS) are
the two types of TLS. To gather high-precision
information, static Terrestrial Laser Scanners (TLS)
involve locating a tripod-mounted laser scanner
within a fixed region and scanning over the
surrounding area using multiple aspects. In
engineering, architecture, and construction, static
TLS is common when exact measurements are
needed for buildings and sites. However, MLS
deploys a handheld scanner like a laser scanner
which is hand-held or even a moving platform such
as a car, drone, backpack, or any other item. MLS
scanners can collect 3D data of the surroundings,
enabling rapid and effective data collection as it
moves around the area. MLS is common among
surveying, mapping, and infrastructure management
applications, [18], [19].
There are benefits and drawbacks to every static
TLS and MLS. When it comes to obtaining more
detailed data for smaller areas, static TLS is always
the best option. It also offers better accuracy and
resolution. On the other hand, MLS is advantageous
for fast and accurate mapping of large regions or
large-scale mapping because it can cover more areas
in a shorter period. Similarly, hazardous, or hard-to-
reach sites such as tall buildings and cliffs can also
be accessed through MLS. By comparing the MLS
technology with static TLS technology, some of its
drawbacks can be noted too. For example, the speed
at which a scan is made may make MLS data less
detailed compared to TLS while other things can
affect how accurate it is also.
Terrestrial laser scanning, or TLS, for static data
capture is vital. It has major benefits like wide
coverage as well as high precision. Nonetheless,
using TLS also has some negative aspects like the
large initial investment needed for equipment and
software, the time-consuming process of capturing a
large or complex project, and the environmental
sensitivity of TLS. To avoid these constraints, a
properly planned TLS survey program can be
implemented, increased using the Reality Capture
(RC) technology’s additional aid during the data
collection procedure, [20].
A well-designed TLS survey plan can save time
on-site, limit self-occlusions, and provide maximum
coverage of aqueduct surfaces with adequate point
density. Typically, TLS is employed in conjunction
with other RC technologies, such as
photogrammetry or SfM (Structure-from-Motion).
By combining TLS with photogrammetry,
practitioners can benefit from both
photogrammetry's flexibility and TLS's high
accuracy, [21].
GoSlam RS100i as shown in Figure 5 was used
to perform Laser scanning measurements. This Slam
laser scanner employs Simultaneous Localization
and Mapping (SLAM) technology which provides
real-time positioning and mapping capabilities.
The GoSlam RS100i can gather 320000 points
per second and has a scanning radius of 120 meters.
Its 360-degree range of view and 1-centimeter point
accuracy make it extremely large.
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.30
Arli Llabani, Otjela Lubonja
E-ISSN: 2224-3496
309
Volume 20, 2024
The data collection for the castle with GoSlam
was 58 min.
Fig. 5: GoSlam RS100i laser scanner
3 Results
3.1 UAV Data Processing
Five processing steps are typically involved in the
generation of a dense, georeferenced 3D point cloud
from a block of overlapping photographs: 1)
Identification of features; 2) matching those
features, among which geometric verification takes
place; 3) SfM sparse 3D reconstruction; 4) GCPs
optimize the bundle adjustment by georeferencing
the scene geometry and adding camera self-
calibration; 5) multi-view stereo dense matching for
dense 3D reconstruction.
There is constant use of the A Scale Invariant
Feature Transform (SIFT) algorithm for detecting
features, i.e., important locations in every image in
the first stage. Important spots are partially invariant
to photometric distortions and 3D camera
viewpoint, and they are scale and rotation
invariance. The texture and resolution of each image
mostly determine how many critical points are
present. In the second stage, an approximate nearest
neighbor (ANN) method is typically used to match
the important locations, which are identified by a
unique descriptor, across many images.
After that, each matched picture pair's key point
correspondences are filtered using a RANdom
SAmple Consensus (RANSAC) algorithm to impose
a geometric epipolar constraint.
The third phase includes reconstructing the 3D
geometry of the scene and the geometry of the
picture network together using iterative bundle
adjustment and geometrically corrected
correspondences (tie points). This further estimates
values that relate to the calibration of a camera—
internal camera (intrinsic) parameters (IOP)—as
well as position and attitude of each given image in
an arbitrary coordinate system—external orientation
(extrinsic) parameters (EOP), with only the tie
points image coordinates acting as observations.
In step 4, the sparse 3D point cloud is scaled
and georeferenced using a seven-parameter 3D
similarity transformation in the GCP coordinate
system. The GCP coordinates are then obtained as
observations and are typically measured using dual-
frequency GNSS receivers. To enhance the EOP,
IOP, and 3D coordinates of the tie points, further
GCP measurements and markings of where they
occur on the images can be included during bundle
adjustment. Rerun the bundle adjustment with
appropriate weights on coordinates of ground (GCP)
and image (tie points and GCP) measurements to
reduce reprojection and georeferencing inaccuracies,
[22].
Stage 5, is often a case of boosting the point
density of the sparse point cloud by several orders
of magnitudes - two or three - using a multi-view
3D surface reconstruction algorithm. From this
reduced general cloud, in which we have both
exterior and interior orientations which were
determined previously as the best one, it performs
the computationally intensive dense matching
technique that creates, initially in image space,
depth maps of every image batch, after which they
are combined into the specified area.
The Pix4DMapper software was used to process
UAV images. This software automatically
transforms the images taken by the drone and
delivers high-precision products such as orthophotos
and Digital Surface Models (DSM). Pix4DMapper
uses the SfM (Structure from Motion) technique to
reconstruct the scene based on many overlapping
photos.
For the processing of aerial images in
Pix4DMapper, initial processing was performed
first. Before beginning the initial processing,
PIX4Dmapper computes the pictures' key points.
This software uses these key points to find the
similarities between the photos. After this initial
match was found, Automatic Aerial Triangulation
(AAT) and Bundle Block Adjustment (BBA) were
then carried out by the software.
In this study, the coordinate system for final
products was chosen UTM Zone 34N (epsg: 32634).
Then, in the Pix4DMapper software, the
corresponding template was defined as 3D maps,
which gives us the products mentioned above.
A digital elevation model (DEM) and an
orthophoto of the surveyed region were produced by
this technique.
The creation of the Point Cloud and 3D mesh
was the second stage of image processing in
Pix4DMapper, following the first processing.
Point clouds are generally produced using 3D
scanners or photogrammetry software that measures
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.30
Arli Llabani, Otjela Lubonja
E-ISSN: 2224-3496
310
Volume 20, 2024
a huge number of points on the external surfaces of
surrounding objects. Point clouds are produced via
3D scanning procedures and are utilized in a wide
range of applications such as mass customization,
animation, visualization, metrology, and 3D
computer-aided design (CAD) models for
manufactured parts.
The model shown in Figure 6 has a high
accuracy with about 303,289 points, 298 pictures,
2.41-pixel size, 5472 x 3648 resolution, and a
camera centering error of 1.2 cm.
After this process, we were able to get the DSM
in 2 cm pixel resolution.
Fig. 6: Point Cloud generated from UAV
Photogrammetry
3.2 Laser Scanning Data Processing
Three primary stages are involved in administering
raw TLS survey scans: (1) register the scan; (2)
clean and optimize point cloud data; and (3) reduce
the point cloud dataset. Scan registration is the first
step in processing scan data. This step will involve
aligning the scans to a single reference system to
generate a point cloud showing the whole heritage
site. Target-based registration, which uses pre-
established targets, or targetless registration, which
uses homologous features, is the usual method for
manually or automatically doing scan registration in
pairs, [23].
Once the scans have been registered, the point
cloud needs to be cleaned to ensure that the final
model is accurate and useful for further research and
documentation. Essentially speaking, point clouds
are collections of three-dimensional data points,
which show the scanned geometry of some
environment or item. However, a variety of factors,
including reflecting surfaces, occlusions, and sensor
noise, might impact these data points, leading to
inaccurate and inconsistent point cloud data.
Cleaning the point cloud means removing any noise
and undesired data points that may have been
gathered, and also correcting any errors that may
have occurred during the scanning process.
To find and eliminate undesirable points, this
procedure usually combines automatic and manual
methods like filtering, segmentation, and
classification.
We used GoSLAM Studio Flagship Version
software to process laser scanning data. This
software integrates point cloud processing and
device application specifically for the GoSLAM line
of mobile 3D scanners. Additionally, it works with
third-party cloud processing devices and point
systems.
The software has eight fundamental features:
coordinate transformation, automatic horizontal
plane fitting, point cloud splicing, forward
photography, automatic point cloud data report
production, one-click point cloud denoising, shadow
rendering, and point cloud encapsulation.
GoSLAM adds one-click heap data generation
to bulk metering to make it easier to access data.
The individually registered TLS point clouds of
the castle were aligned using GoSLAM Studio
Flagship. We combined all the data after this
alignment to create a single point cloud as shown in
Figure 7.
A collection of points, similar to pixels in a
digital image, is called a point cloud. While the
points of the point cloud are made up of three
coordinates—X, Y, and Z—each of which
represents a distinct location in the three-
dimensional space, each pixel is made up of two
coordinates, X and Y. Millions of points come
together to form a 3D shape or view in a point
cloud, [24].
The maximum point error was 2.8 mm and the
mean point error was 2.1 mm.
Fig. 7: Point Cloud generated from Terrestrial Laser
Scanning
3.3 Comparison between UAV and Laser
Scanning
This section contains the findings of our comparison
of accuracy in the chosen field using two different
methodologies. Based on the results shown, GoSlam
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.30
Arli Llabani, Otjela Lubonja
E-ISSN: 2224-3496
311
Volume 20, 2024
laser scanning is more accurate than UAV
photogrammetry when it comes to documenting
cultural heritage sites, particularly castles.
The point clouds produced by these techniques
were compared using Cloud Compare software.
After importing the clouds into CloudCompare, we
were able to determine their separation from one
another. The point cloud that was selected as a
reference came from measurements made with a
laser scanner.
The largest distance between these point clouds
according to Cloud Compare is 67.727 meters and
the average distance is 2.095 m.
Fig. 8: Distance Computation in Cloud Compare
Figure 8 illustrates how errors and noise can be
found in the point cloud produced by UAV
photogrammetry. The point cloud acquired using
laser scanning has better geometric accuracy.
We obtained a mean RMS of 2.199073 mm and
std. dev was 7.356 mm.
By taking advantage of one technology's
benefits over the other, TLS and photogrammetry
can work together. Compared to photogrammetric
surveys, TLS is a more costly technology, and
performing laser scans requires more expertise than
taking photos for 3D photo reconstructions.
However, photogrammetry is more efficient,
adaptable, quick, and able to gather high-quality,
precise data for intricate things. Combining these
two technologies allows for the utilization of the
TLS's precise and dense point cloud acquisition
capabilities and photogrammetry's adaptability to
work under extreme circumstances.
It has been shown that the best approach to
recording big and complex heritage sites for
purposes such as documentation, structure
evaluation, texture mapping, feature extraction, etc.
is to combine TLS with photogrammetric
approaches.
4 Discussion
This study investigated the combination of Aerial
Photogrammetry and TLS for conducting a 3D
survey of the Fortress of Bashtova. TLS and UAV
photogrammetry are best understood as aggregated
because of the differences in data quality between
them.
Thus, by obtaining the point cloud data along
with their attributes, we were able to recognize risks
in the heritage site documentation. In conclusion,
this gives the approach a perfect starting point for
the documentation of heritage areas. We used
CloudCompare to perform an accuracy analysis of
each remote sensing point cloud to address the
significant issue that was brought up. As a result, we
attempted to compare the point clouds produced by
TLS and UAV.
After analyzing our results, we found that there
is a 1.7-centimeter difference between the point
clouds obtained from TLS and UAV.
Our study presents a workable framework for
the integration of TLS and UAV-based
photogrammetry with applications to heritage areas.
This framework includes TLS and UAV image
acquisition, point cloud processing, and 3D mapping
of heritage areas. Indeed, UAV photogrammetry has
been widely used in a variety of fields, with some
work in heritage areas documentation having been
recorded.
Our study's findings show that unmanned aerial
vehicles (UAVs), which require authorization to fly,
may be deployed swiftly and often, can fly at low
altitudes with less cloud interference, and are less
expensive than manned planes and satellites.
Nevertheless, there are a few drawbacks to the
suggested strategy. First, it can occasionally be
challenging to locate ground-characteristic features
in historical places. The second is that to cover a
greater area, a UAV campaign needs more fly routes
due to the length of the battery. Due to the aerial
photos' angle of view and distance, the UAV's point
cloud does not have information on the fortress's
entrance. However, it provides a comprehensive
model of the fortress's uppermost section. The TLS
point cloud makes it possible to more precisely and
in detail represent the entrances and facades.
5 Conclusion
To effectively protect and research cultural heritage
locations, aerial photogrammetry and terrestrial
laser scanning offer various advantages that should
best be chosen or combined based on the required
level of accuracy, data coverage, and visual detail. It
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.30
Arli Llabani, Otjela Lubonja
E-ISSN: 2224-3496
312
Volume 20, 2024
should be highlighted that while researching cultural
heritage sites, high-resolution methods like TLS and
UAV photogrammetry are the best options because
they make it possible for us to gather accurate data
for the documentation of heritage sites.
Laser scanning technology consistently delivers
higher spatial accuracy due to direct point
measurements. After analyzing our results, the
maximum point error of the point cloud derived
from TLS was 2.8 mm and the mean point error was
2.1 mm. This makes it the preferred choice when
precise dimensional information is critical, such as
for intricate architectural elements. It is also
economical and perfect for the detailed scanning of
objects.
UAV photogrammetry excels in providing
broad coverage efficiently, making it suitable for
documenting larger cultural heritage areas. Its aerial
perspective can capture extensive terrains and
architectural layouts swiftly.
According to the study's findings, the TLS
offers a more accurate accuracy than UAV
Photogrammetry for the documentation of cultural
heritage sites. Results showed that there is a 1.7-
centimeter difference between the point clouds
obtained from TLS and UAV. By comparing the
point clouds derived from TLS and UAV in
CloudCompare software, we obtained a mean RMS
of 2.199 mm and std. dev was 7.356 mm.
While the accuracy gained via UAV
photogrammetry was 2 cm, the accuracy obtained
from the Terrestrial Laser scanning approach for
creating the point cloud was 2.8 mm. It should be
noted that the TLS will not deliver data from the
upper parts of the castle if it is used as a device
carried by the operator. Therefore, it is always
advantageous to use a combination of aerial
photogrammetry and TLS to create a comprehensive
model of the object.
This study offers a useful and simple research
proposal for deterioration analysis, TLS
measurements, and UAV photogrammetry in
addition to 3D modeling. It is evident from the
research and findings provided in this article that the
approach discussed in the article is appropriate for
architectural and conservation investigations.
In general, we can recommend both
technologies for the documentation of the heritage
sites.
It is crucial to remember that the UAV approach
might need more specialist tools and knowledge,
and it might be impacted by things like air
anomalies or poor image processing. UAV
photogrammetry is more advantageous than
Terrestrial laser scanning in terms of cost and time
since it takes less time to photograph the region that
needs to be measured.
Declaration of Generative AI and AI-assisted
Technologies in the Writing Process
During the preparation of this work, the authors
used ChatGPT to enhance the clarity and coherence
of the text. After using this tool/service, the authors
reviewed and edited the content as needed and take
full responsibility for the content of the publication.
References:
[1] Albertz J., “A Look Back: 140 Years of
Photogrammetry: In Photogrammetric
Engineering & Remote Sensing, ‘‘ vol. 73, no.
5, pp. 504-506, 2007.
[2] Akgul M., Yurtseven M., Gulci S., Akay A.E.,
“Evaluation of UAV and GNSS-Based DEMs
for Earthwork Volume,’’ Arabian Journal for
Science and Engineering, vol. 43, no. 4, pp.
1893-1909, 2018. DOI: 10.1007/s13369-017-
2811-9.
[3] Balázsik V., Tóth Z., Abdurahmanov I.,
“Analysis of Data Acquisition Accuracy with
UAV,” International Journal of
Geoinformatics, vol. 17, no. 1, pp. 1-10, 2021.
https://doi.org/10.52939/ijg.v17i1.1697.
[4] Erdelyi J., Kopacik A., Kyrinovic P.,
“Construction control and documentation of
facade elements using terrestrial laser
scanning” Applied Geomatics, vol. 10, no. 2,
pp. 113-121, 2018. DOI: 10.1007/s12518-
018-0208-4.
[5] Guo M., Sun M., Pan D., Wang G., Zhou Y.,
Yan B., Fu Z., “High-precision deformation
analysis of yingxian wooden pagoda based on
UAV image and terrestrial LiDAR point
cloud” in Heritage Science, vol. 11, no. 1, pp.
1-18, 2023. DOI: 10.1186/s40494-022-00833-
z.
[6] Haala N., Alshawabkeh Y., “Combining Laser
Scanning and Photogrammetry—A Hybrid
Approach for Heritage Documentation. In M.
Ioannides, D. Arnold, F. Niccolucci, & K.
Mania (Eds.) ”, The 7th International
Conference on Virtual Reality, Archaeology
and Intelligent Cultural Heritage, Nicosia,
Cyprus 2006, pp. 163-170. DOI:
10.2312/VAST/VAST06/163-170.
[7] Hassan A.T., Fritsch D., “Integration of Laser
Scanning and Photogrammetry in 3D/4D
Cultural Heritage Preservation—A Review.‘‘
International Journal of Applied Science and
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.30
Arli Llabani, Otjela Lubonja
E-ISSN: 2224-3496
313
Volume 20, 2024
Technology”, vol.9, no.4, pp. 9-16, 2019,
DOI: 10.30845/ijast.v9n4p9.
[8] Jaafar H.A., Meng X., Sowter A., Bryan P.,
“New approach for monitoring historic and
heritage buildings: Using terrestrial laser
scanning and generalized Procrustes
analysis.” in Structural Control and Health
Monitoring, Vol. 24, no. 11, DOI:
10.1002/stc.1987.
[9] Pavelka K., Šedina J., Pacina J., Plánka
L.,Karas J., Šafář V. “RPAS Remotely Piloted
Aircraft System”, in České vysoké učení
technické v Praze, Prague, Czech Republic,
ISBN: 978-80-01-05648-6, 2016.
[10] Kwoczynska B., Piech I., Polewany P., Gora
K., “Modeling of sacral objects made based
on aerial and terrestrial laser scanning”, in
Baltic Geodetic Congress, Olsztyn, Poland,
2018, pp. 275-282. DOI: 10.1109/BGC-
Geomatics.2018.00059.
[11] Logothetis S., Delinasiou A., Stylianidis E.,
“Building information modelling for cultural
heritage: a review”, In ISPRS Annals of
Photogrammetry, Remote Sensing and Spatial
Information Sciences, Vol. II-5/W3, pp. 177-
183, 2015. DOI: 10.5194/isprsannals-ii-5-w3-
177-2015.
[12] Matrice 300 RTK User Manual, 2020,
[Online].
https://dl.djicdn.com/downloads/matrice-
300/20200507/M300_RTK_User_Manual_E
N.pdf (Accessed Date: November 2, 2023).
[13] Mohammadi M., Rashidi M., Mousavi V.,
Karami A., Yu Y., Samali B., “Quality
evaluation of digital twins generated based on
UAV photogrammetry and TLS: bridge case
study”, in Remote Sensing, vol.13, no. 17,
pp.3499, 2021. DOI: 10.3390/rs13173499.
[14] Moon D., Chung S., Kwon S., Seo J., Shin J.,
“Comparison and utilization of point cloud
generated from photogrammetry and laser
scanning: 3D world model for smart heavy
equipment planning”, in Automation in
Construction‘‘, vol. 98, pp. 322-331, 2019.
DOI: 10.1016/j.autcon.2018.07.020.
[15] Mulakala J. “Measurement Accuracy of the
DJI Phantom 4 RTK & Photogrammetry”,
[Online]. https://www.gim-
international.com/files/23b0ad77f81a0aa56e8
c83f8c4300270.pdf (Accessed Date: October
28, 2023).
[16] Przybilla H.J., Baeumker M., “RTK and PPK:
GNSS-Technologies for Direct
Georeferencing of UAV Image Flights”, FIG
Working, Week Conference: FIG Working
Week 2020 At: Amsterdam, 2020, pp. 2-16,
[Online].
https://www.fig.net/resources/proceedings/fig
_proceedings/fig2020/papers/ts01b/TS01B_pr
zybilla_manfred_10801_abs.pdf (Accessed
Date: November 14, 2023).
[17] Pritchard, D., Sperner, J., Hoepner, S., and
Tenschert, R.: Terrestrial laser scanning for
heritage conservation: the Cologne Cathedral
documentation project, ISPRS Ann.
Photogramm. Remote Sens. Spatial Inf. Sci.,
IV-2/W2, 2017, pp. 213–220,
https://doi.org/10.5194/isprs-annals-IV-2-W2-
213-2017.
[18] Pix4D. “Pix4Dmapper User Manual.
Lausanne”, 2020, [Online].
https://s3.amazonaws.com/mics.pix4d.com/K
B/Getting+Started+PDFs/EN/Pix4Dmapper+-
+Getting+Started+-+Master+-+4.0+-+EN.pdf
(Accessed Date: November 6, 2023).
[19] Pompejano F. “From national to cultural
monuments: an overview of architectural
heritage protection in Albania (1912-1992),”
Journal of Architectural Conservation, vol.26,
no. 1, pp. 55-70, 2020. DOI:
10.1080/13556207.2019.1684021.
[20] Remondino F., “Heritage recording and 3D
modeling with photogrammetry and 3D
scanning”, in Remote Sensing, vol. 3, no. 6,
pp. 1104-1138, 2011. DOI:
10.3390/rs3061104.
[21] Rodríguez-Gonzálvez P., Jiménez Fernández-
Palacios B., Muñoz-Nieto Á.L., Arias-
Sanchez P., Gonzalez-Aguilera D. “Mobile
LiDAR System: New Possibilities for the
Documentation and Dissemination of Large
Cultural Heritage Sites”, Remote Sensing, vol.
9, no. 2, pp.189, 2017. DOI:
10.3390/rs9030189.
[22] Uysal M., Toprak A.S., Polat N., “DEM
Generation with UAV Photogrammetry and
Accuracy Analysis in Sahitler Hill.
Measurement”, vol. 73, pp. 539-543. 2015.
https://doi.org/10.1016/j.measurement.2015.0
6.010.
[23] Yastikli N., “Documentation of cultural
heritage using digital photogrammetry and
laser scanning”, Journal of Cultural Heritage,
vol. 8, no. 4, pp. 423-427, 2007.
https://doi.org/10.1016/j.culher.2007.06.003.
[24] Wang Y., Chen Q., Zhu L., Liu L., Zheng, L.,
“Survey of Mobile Laser Scanning
Applications and Key Techniques over Urban
Areas” in Remote Sensing, vol. 11, no. 13, pp.
1540, 2019. DOI: 10.3390/rs11131540.
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.30
Arli Llabani, Otjela Lubonja
E-ISSN: 2224-3496
314
Volume 20, 2024
Contribution of Individual Authors to the
Creation of a Scientific Article (Ghostwriting
Policy)
- Arli Llabani carried out the UAV and TLS
measurements and their comparison using
CloudCompare software.
- Otjela Lubonja was responsible for the
architectural analysis of the Fortress of Bashtova.
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 conflicts of interest to declare.
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
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.30
Arli Llabani, Otjela Lubonja
E-ISSN: 2224-3496
315
Volume 20, 2024
