Generating three dimensional photo-realistic model of archaeological monument via multi-sensors’ data integration

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
__________________________________________________________________________________________
Volume: 02 Issue: 10 | Oct-2013, Available @ http://www.ijret.org 362
GENERATING THREE-DIMENSIONAL PHOTO-REALISTIC MODEL OF
ARCHAEOLOGICAL MONUMENT VIA MULTI-SENSORS’ DATA
INTEGRATION
Nurul Shahida Sulaiman1
, Erna Warnita Bachad2
, Albert K. Chong3
, Zulkepli Majid4
, Halim Setan5
1, 2, 4, 5
Faculty of Geo information and Real Estate, Universiti Teknologi Malaysia, Johor, Malaysia
3
Schools of Civil Engineering and Surveying, University of Southern Queensland, Australia,
shahida.sulaiman@gmail.com, ernawarnita@gmail.com, chonga@usq.edu.au, drzulmajid@gmail.com, halim@utm.my
Abstract
Stimulated by the growing demand for three-dimensional (3D) photo-realistic model of archaeological monument, people are looking
for efficient and more precise methods to generate them. However, a single method is not yet available to give adequate results in all
situations, especially high geometric accuracy, automation, photorealism as well as flexibility and efficiency. This paper describes the
method for generating 3D photo realistic model of Bukit Batu Pahat shrine by the means of multi-sensors data integration. Global
Positioning System (GPS), Total Station and terrestrial laser scanner (Leica ScanStation C10) were used to record spatial data of the
shrine. Nevertheless, due to the low quality of the coloured point cloud captured by the scanner, a digital camera, Nikon D300s was
used to capture photos of the shrine for surface texturing purpose. A photo realistic 3D model with high geometric accuracy of the
shrine was generated through spatial and image data integration. A feature mapping accuracy and geometric mapping accuracy was
conducted to analyse the quality of the 3D photo-realistic model of the monument. Based on the results obtained, the integration of
ScanStation C10 and images data is capable of providing a 3D photo-realistic model of Bukit Batu Pahat shrine where feature
mapping accuracy showed that the model was 90.56 percent similar with the real object. Additionally, the geometrical accuracy of the
model generated by ScanStation C10 data was very convincing which was ±4 millimetres. In summary, the goal of this research has
successfully achieved where a 3D photo- realistic model of Bukit Batu Pahat shrine with good geometry was generated through multi-
sensors data integration.
Keywords: archaeological monument, terrestrial laser scanner, digital photo, 3D photo-realistic model
-----------------------------------------------------------------------***-----------------------------------------------------------------------
1. INTRODUCTION
In the past two decades, three-dimensional (3D) modelling
was becoming a prevalent trend in many researches. Instead of
focusing on the typical areas of applications such as car
design, reverse engineering for precision mechanics or
medical imaging, 3D modelling is becoming popular in
various fields like anthropometrics, archaeology and many
others (Guarnieri et al., 1999). According to Aguilera and
Lahoz (2006), the demand for 3D models of historical
monuments is increasing in the field of archaeological and
architectural applications.
Several techniques and tools were developed to fulfil the
increasing demand of 3D model especially in designing and
manufacturing industries. Nonetheless, as stated by Campana
et al. (2009), in order to conduct archaeological research, it is
crucial to utilize Geomatic techniques because these
techniques are capable in representing correctly all
peculiarities of the investigated object. Fig-1 shows the
capability of two Geomatic tools; laser scanner and
photogrammetry in recording the inscription on the wall. Both
approaches resulted with same accuracy which is 0.3
millimetres respectively.
Fig-1: Geometric details of an inscription captured using
different sensors (left: laser scanner, right: photogrammetry)
(Remondino, 2010)
Basically, a systematic documentation in archaeological
research is important for further analysis and interpretation.
Nevertheless, a digital documentation is more preferable

__________________________________________________________________________________________
because analysis and interpretation can be done virtually
without worrying about access limitation or physical damage
of the remains. Hence, image-based method (i.e.
photogrammetry), range-based method (i.e. laser scanning)
and integration between the two methods are suitable
approaches to be used as they are non-contact methods that
prevent handling damage of archaeological remains,
applicable in various scales and yet the sensors, processing
software and analysis tools have become available in recent
years (Lambers and Remondino, 2007).
In this research, the archaeological area is approximately 48 x
40 metres which contains hundreds of religious artefacts and a
temple or shrine. In order to generate highly accurate
geometric and photo-realistic model, several Geomatic tools
were used to record spatial and texture data of the shrine.
Terrestrial laser scanner, Leica ScanStation C10, was used to
collect spatial information of the structure while Nikon D300s
digital camera was used to capture high resolution images.
Network control for georeferencing purpose was established
around the area by using Global Positioning System (GPS)
and total station. A modelling methodology was carried out by
integrating data from all the sensors.
2. RELATED WORKS
In archaeology, a systematic and well-judged use of Geomatic
techniques for documentation and digital conservation
purposes is relatively a recent innovation, not yet applied on a
regular basis for a variety of reasons; the perceived „high cost‟
of capturing 3D, the difficulties in achieving good 3D models,
the difficulty of integrating 3D worlds with other 2D data and
traditional documentations and the perception that 3D is only
an additional „aesthetic‟ factor (Campana et al., 2008).
However, the available technologies and methodologies for
digital recording of archaeological sites and objects are really
promising and the whole heritage community is trying to adapt
these approaches for fastest, detailed and easy 3D
documentation (Campana and Remondino, 2009).
Methods and techniques chosen for archaeological monument
documentation should be the most effective, least destructive,
most efficient and economical means of obtaining the needed
information. In the beginning, there was only a short range
laser scanner and mainly used in the automotive and industrial
design process to facilitate the data capture for Computer
Aided Design (CAD) process. Nevertheless, since technology
keeps evolving, other potential fields were exploited. Due to
outstanding advantages of laser scanning; non contact
measurement, high accuracy, long range, fast data acquisition,
etc., other disciplines like cultural heritage, architecture, urban
development, forensics, and the entertainment industry were
starting to adopt this technology (Heine, 2004).
When dealing with archaeological structure documentation,
end user normally require a geometric documentation of the
monument as a final product. Primarily, geometric
documentation of archaeology monument is very important
because it consists of a series of measurements, from which
visual products such as vector drawings, raster images, 3D
visualizations etc. may be produced (Georgopoulos and
Ioannidis, 2004). Georgopoulos and Ioannidis (2004) also
stressed that the products‟ scale, purpose and accuracy
requirement must be defined first before starting the
documentation process (Table-1). This is because the scale of
the final product implies the level of detail yield eventually.
Table-1: The mapping scale requirement for geometric
documentation (Georgopoulos and Ioannidis, 2004)
A geometric documentation of a monument contains
measurement of the objects which will be used for restoration
or reconstruction work of the monument. The selection of
recording type will be based on the accuracy demanded by the
client. More accurate geometric documentation of the
monument will require lots of effort and the process of
collecting and producing the output will be more complicated.
Table-2 shows the accuracy requirement needed in
archaeological site recording.
Table-2: Accuracy requirement in archaeological site
recording (Georgopoulos and Ioannidis, 2004)
Terrestrial laser scanning technology is becoming increasingly
popular recently as the scanning devices become more
reliable, cheaper, faster and more portable (Fang et al., 2009).
By taking into consideration about the size of the objects, the
distance between object and scanner, the surface of the object
Geometric
Documentation Scale
Purpose
1:100 – 1:250 General surveys, in order to
relate the monument to its
immediate surroundings
1:20 – 1:50 Geometric recording cases
providing a highly detailed
product for practically any sort
of study
less than 1:20 Drawings of special details of
interest
Type of
Recording
Purpose Accuracy
Requirement
Normal
Recording
Structural / Building /
Others
± 2cm
Plans / Elevation / Cross
Section
± 10cm
Detailed
Recording
Building / monument ±2mm to ±5mm
Building Plans / Elevation
/ Cross Section
± 10mm to
±25mm

__________________________________________________________________________________________
and the accuracy requirement, a suitable type of range scanner
is chosen. Lambers and Remondino (2007) stated that there
are three scales of archaeological research where Geomatic
techniques and methodologies can be applied. Table-3 shows
the detail on the scale of archaeological site and the suitable
Geomatic tool to be implemented (Lambers and Remondino,
2007), (Guidi et al., 2008), (Abdullah, 2005).
Table-3: Size of archaeology site versus suitable Geomatic
tools (modified from Lambers and Remondino, 2007)
During conventional recordings, different people often have to
draw the features and the results are non-homogeneous
drawings. Basically, the accuracy for manual drawing is 25
millimetres which required a lot of time, focus and skills in the
documentation process. However, the accuracy requirement
for geometric documentation of archaeology monument is 70
millimetres to few millimetres (Georgopoulos and Ioannidis,
2004). Additionally, a laser scanner can achieve up to 2.00
millimetres accuracy when the measuring range is less than
120 metres with 50000 points per second.
The need of combining multiple techniques, like terrestrial
laser scanning, photogrammetry and digital surveying is the
result of the complexity of some structures and by the lack of
a single technique capable of giving satisfactory results in all
measuring conditions (Gonzo et al., 2007). There is no single
method that is applicable of recording every subject of cultural
heritage and hence there is a strong demand for a hybrid
method that exploits several technologies (Kadobayashi et al.,
2004). The advantage of using laser scanning is that it can
produce a very dense 3D point cloud data which is a crucial
requirement for creating high-resolution geometric models
(Fig-2). On the other hand, digital photogrammetry can
produce high-resolution texture if the original images have
sufficient resolution (Fig-2). Thus, the integration of these two
technologies is indeed a key to create high quality 3D
recordings and presentations.
Fig-2: Porta' de Santiago; Image (left). 3D textured model
(Ong et al., 2010)
3. SITE DESCRIPTION AND CURRENT
DOCUMENTATION IN BUJANG VALLEY
Being situated at the southern tip of Asian mainland,
Peninsular Malaysia which is also located between the major
trades centres of China in the East and India and Arab world
in the west seems to contribute enormously in the maritime
trade. Thus, it possesses some remarkable archaeology sites
that are significant and require preservation such as in
Lenggong Valley and Bujang Valley.
Bujang Valleys is the home of an impressive, well-preserved
historical complex that is capable of eliciting awe and
wonderment for those who come to visit the ancient site. To
date, more than 80 archaeology sites have been discovered in
the valley (New Straits Times, 2008). According to Nik
Hassan Shuhami (2004), among of all excavated sites, 60 sites
were thought to have been religious shrines. The most
prominent shrines discovered in the valley are Bukit Batu
Pahat shrine (Site 8), Bukit Choras shrine (Site 1), Bendang
Dalam shrine (Site 50) and Pengkalan Bujang shrine (Site 21).
Nevertheless, archaeological activities in Malaysia are
progressing slowly due to the lack of expertise in
archaeological field, the use of conservative technique for
documentation purpose (Fig-3) and less awareness among
Malaysian on the important of preserving the remains.
Archaeology and conservation activity in Malaysia has started
in early 1990s and it is far left behind compared to other
develop countries which have started these activities since
1960s (Nik Hassan Shuhaimi, 2006).
Size of
Archaeo
logy Site
Object of
Interest
Available
Geomatic
Approaches
Geometric
Resolution
(per pixel)
Regional Landscape
Topograph
y
Sites /
Monument
-Middle and high-
resolution
satellite imagery
-Small scale aerial
images
-Airborne LiDAR
IKONOS: 1 m
QUICKBIRD:
0.6m
0.50 m - 0.75
m
0.75 m - 3.0 m
Local Sites /
Monument
Architectur
es
Excavation
Layers
-Large scale aerial
images
-Terrestrial laser
scanner
-Terrestrial
images.
-10 cm- 25 cm
-5 mm - 20
mm
-0.5 mm - 10.0
mm
Object Excavated
artefacts
Museum
objects
-Terrestrial
images.
-Close-range laser
scanner
-0.5 mm - 10.0
mm
-0.3 mm

__________________________________________________________________________________________
Fig-3: Conventional methods used in Malaysia archaeological
research (Patel, 2009)
An analogue filing system is used by the administration which
seems to be impractical due to the increasing excavation
activities, require bigger storage space, redundancy of
information forms for artefacts and misinterpretation of
handwriting. Hence, it is a crucial requirement for researches
to develop a digital documentation system to improve the
existing documentation method in Malaysia. Thus, this
research is a pilot project in implementing new method for
providing digital data of archaeological site in Bujang Valley.
An archaeological site, Bukit Batu Pahat shrine (Fig-4) which
has been excavated was chosen as a study area where the
shrine is modelled into 3D photo-realistic model.
Fig-4: Location of the study area; Bujang Valley and Bukit
Batu Pahat shrine
4. METHODOLOGY
Numerous methods were developed to generate 3D model of
archaeological monument. Unfortunately, most contributions
were focusing either on a single sensor i.e. camera or
integration between range sensor and image sensor.
Conversely, this paper attempts to reveal the integration
process involving various sensors with different spatial
resolution and types of data. Multiple sensors were used in site
survey and archaeological monument mapping phase. Site
survey provides coordinate information of the monument with
respect to the real terrain world while archaeological
monument mapping provides information such as geometrical
characteristic and texture of the structure.
Based on Fig-5, this research focused with two activities that
involved in documenting archaeological monument; site
survey and monument mapping. Site survey is a task where
control points with known mapping coordinates are
established around the archaeological site. GPS static
observation was conducted where two points were observed.
As to the production of a precise model, a control network was
established by using a total station. Coordinate yield from site
survey task is used to georeference the model.
In monument mapping phase, two sensors were used to collect
spatial and texture data of the shrine. Leica ScanStation C10
(terrestrial laser scanner) was used to collect spatial data of the
shrine. The data collected were in the form of 3D point cloud.
In the mean time, the scanner also captured images using a
built-in camera inside the scanner itself. It has been proven in
many researches that images captured from a built-in camera
in a scanner provide only low quality of images and thus
affecting the result for photo-realistic model especially for an
outdoor task. Hence, a high resolution digital camera was used
to capture images of the shrine. The images were then fused
with the laser scanner data to get a 3D photo-realistic model of
the monument with good geometrical accuracy.

__________________________________________________________________________________________
Site Survey
Global Positioning
System (GPS)
Archaeological Site
Monument
Monument Mapping
Total Station
Terrestrial
Laser Scanner
Digital Camera
Coordinate
data (.txt)
Traverse
data
Point Cloud
data (.bin)
Digital
images (.jpg)
Point Cloud
Processing (Leica
Cyclone)
Point cloud registration
and alignment using
Cyclone
Giving GPS coordinates
to all traversing points
3D Mesh
Model (.rsh)
Texturing the
Mesh (3D
Reshaper 7.1)
3D Photo-realistic
Model (.rsh)
Photo-realistic Model
Evaluation
(Feature Mapping Accuracy)
3D Models Visualization in
3D Reshaper Viewer
DataCollection
RawDataProcessingand3D
Modelling
3DModelEvaluationand
Visualization
Creating 3D Mesh
(3D Reshaper 7.1)
3D Point Cloud
Model (.xyz)
Georeferencing
Selected
Digital
Images (.jpg)
Mesh Model Evaluation
(Geometric Mapping
Accuracy)
Combine point cloud
data from Faro scanner
(.fls)
Fig-5: Workflow to generate 3D photo-realistic model of
Bukit Batu Pahat shrine
4.1 Data Acquisition
4.1.1 Site Survey
Site survey at Bukit Batu Pahat shrine was conducted using
GPS and a Total Station. Two points were observed using
GPS with static mode within an hour (Fig-6). The coordinates
of the two points were referred to Malaysian Cassini State
Plane coordinate. The same points were then used during
traversing process. Four points were established during
traversing around the shrine. The points were marked
permanently because the same points are needed to locate the
scanner during scanning process.
Fig-6: Location of GPS points, traverse points and respective
coordinates for each point
4.1.2 Scanning the Monument
The second phase in recording archaeological monument is to
map spatial information and features of the monument.
Therefore, the monument mapping phase is divided into two
parts; spatial data acquisition using terrestrial laser scanner
and features data acquisition using high-resolution digital
camera. Fundamentally, before a structure is scanned, the area
must be observed first to plan the location of scanning stations
and to decide suitable places to put the target. If the targets are
not enough, user may also used checkerboard target that can
be directly printed from Cyclone software. Before scanning is
conducted, all the targets were placed covering the object
where at least three targets can be seen from two consecutive
stations.
Direct georeferencing method was used where the scanner was
setup at the exact points established during the traversing
process. This is because; the points have coordinates observed
using GPS. Hence, traversing and scanning using terrestrial
laser scanner can be carried out using the same reference
frame. The requirement for spatial information of the
monument is that the data is in the form of digital information.
Thus, the data can be used for various purposes such as
analysis by archaeologists, study by researches, archive by
preservation and conservation department and used in tourism
activities. Consequently, the manipulation of the data can be
carried out while the existing monument in the site is
preserved.
A terrestrial laser scanner, Leica ScanStation C10 was used to
record 3D spatial information of Bukit Batu Pahat shrine.
Initially, a survey planning is made by defining the purpose of
mapping the monument, accuracy demanded and determining
the deliverables as required by the clients. This was followed
by a site observation task to know the complexity of the
structure, number of scanning station and target location. After
defining the requirement and complexity of the structure, the
method or equipment to be used in the data acquisition is

__________________________________________________________________________________________
confirmed. In this research, the required deliverable is a 3D
photo-realistic model of Bukit Batu Pahat shrine with accurate
geometric model.
The project name, resolution of scanning, mode of scanning
either scan with image or just scan, space available in the
storage and method of scanning (traverse and scanning,
scanning only) must be verified first. For Bukit Batu Pahat
shrine, the resolution used was high resolution (0.05 meter of
point spacing), method of scanning was traverse and scan and
mode of scanning was scan with image. Scanning with high
resolution would take about forty five minutes for each scan
station. Each HDS target was scanned before scanning the
whole site. Even though it seems difficult, but this step could
reduce the processing time later. Images were captured by a
built-in camera (4 megapixels) inside the scanner after each
scanning finished. Scanning process was continued until the
area of the shrine was fully covered. Fig-7 illustrates the
position of scanning stations and targets distribution around
Bukit Batu Pahat shrine.
Fig-7: Target distribution and location of scanning stations
4.1.3 Capturing High-Resolution Images
Museums are keen on presenting their collection in a more
appealing and exciting manner using the digital media to
attract visitors to view virtually and visit the physical museum
site. According to White et al. (2004), current surveys show
that about 35% of museums have already started
developments with some form of 3D presentation of objects.
Development of 3D laser scanner has brought a whole new
landscape to the recording and study of historical and cultural
heritage. Nevertheless, the drawback of laser scanner is that
colour images of an object cannot be obtained or even it is
possible, the quality of the colour images are worst than that
of the photographs (Kodabayashi et al., 2004). Fig-8 shows
coloured point cloud generated by Leica Scan Station C10.
Hence, digital photographs were taken by using Nikon DSLR
D300s with 10.2 megapixels.
Fig-8: Coloured point cloud of Bukit Batu Pahat shrine
provided by Lieca Scan Station C10
Instead of providing colour information for the scanning
points, digital images have also been used to extract
information on edges and linear surface features in order to
bridge gaps in the laser scanner data and to add new details to
improve the realistic perception of the scene (Alshawabkeh
and Haala, 2004). If a direct photo-shots texture mapping is
used for texturing the 3D point clouds, the photographs must
have been captured by a high resolution camera located
perpendicular with the object (Al – Hanbali and Awamleh,
2010). The high-resolution images were used for texturing the
3D model to get a better photo-realistic 3D model.
4.2 Data Processing
The main goal of this research was to integrate data from
various sensors in order to generate 3D photo-realistic model
of Bukit Batu Pahat shrine. Nevertheless, data from each
sensor must be processed first before they can be integrated.
Thus, there were three data processing tasks involved in this
research; point cloud registration, point cloud modelling and
texturing the 3D model.
4.2.1 Point Cloud Registration
Registration is the process of merging every ScanWorld
contained in a project into a single registered ScanWorld. The
process of registering the ScanWorlds was carried out by
systematically merges two contiguous ScanWorlds via
corresponding constraints or corresponding point cloud
appeared in both ScanWorlds. There were two methods of
registration provided by Cyclone 7.3; (1) combining the
ScanWorlds via corresponding target or constraint such as
HDS target or black and white target, (2) combining the
ScanWorlds via cloud-to-cloud method where at least three
corresponding points from each ScanWorlds was chosen as
common points to merge two ScanWorlds. Eventually, all the
ScanWorlds were merged together to form a complete 3D
point cloud model of Bukit Batu Pahat shrine.

__________________________________________________________________________________________
Generally, more constraints or corresponding points marked
could give better result. However, Cyclone requires only three
corresponding points to merge two ScanWorlds. Once the
points were selected, Constraint menu was chosen from the
Cloud Constraint wizard to add the constraints into the
database. Error might occur whereby the constraint was not
accepted. In this issue, any bad point was removed and other
point is selected. In the Optimize Constraint stage the
constraints were analyzed whether they could be used for the
registration or vice versa. Fig-9 shows the scanning data sets
that acquired from different scanning points before and after
the registration process.
Before registration
After registration
Fig-9: Bukit Batu Pahat shrine; before and after registration
4.2.2 Point Cloud Modelling
A point-based representation of an object can give satisfactory
result only when the object is viewed in a small scale. But
when viewing in a large scale, the point cloud model started to
deteriorate and what can be seen is just points scattered
randomly. Therefore, the point clouds are modelled into a
surface mesh, which best represent the reality. A mesh model
is a collection of contiguous polygons that combine along their
edges. In this research, a mesh model of Bukit Batu Pahat
monument is indeed the basis for various 3D model
generations such as 3D photo-realistic and 3D geometric
model of the structure. The process of meshing the point cloud
model was carried out using 3D Reshaper 7.1 software.
Once a mesh model of the shrine was produced, the results
were analyzed based on the triangles that covered holes in the
model, the existing of anomalous points and the edges of the
structure. If the result was poor, the point cloud model was re-
processed by changing the parameters value. Two critical
parameters were considered when creating a mesh using 3D
Reshaper 7.1; average distance between points and triangle
sizes. Smaller value for average distance between points could
produce a surface like the real surface because the points were
denser and more triangles were created. In addition, for any
small holes detected, the software automatically detected and
filled the holes with triangles. Thus, smaller triangles could
increase the sharpness of edges.
4.2.3 Texturing the 3D Mesh Model
Although the most important aspect in documenting an
archaeological monument is the geometry of the structure,
preserving its texture is also essential because a photorealistic
model can be used for virtual advertising and further
conservation purpose (Aries et al., 2005). Concern on the
arising demand in archaeological documentation, this research
was carried out to find a method to deliver both metric and
textured 3D model of archaeological monument. Moreover,
instead of providing method to generate 3D photo-realistic
model of archaeological remains, this research was also
intended to find an alternative for better digital visualization,
preservation method and documentation of archaeological
remains. By having a quality 3D textured model, it could
improve the effectiveness of data dissemination among
archaeologists, researchers, historians, students and tourists.
In this research, Reference point’s method was chosen to map
digital images onto the 3D mesh model. It is suggested to
mark at least four pairs of corresponding points if the image
has perspective projection (Fig-10). Nevertheless, higher
number of points is essential for image that has distortions.
When all the points defined and marked correctly, the image
can be applied onto the 3D mesh model. Fig-11 shows the
result before and after applying digital image onto 3D mesh
model.

__________________________________________________________________________________________
Fig-10: Corresponding points are marked on the 3D mesh
model and images
Fig-11: Mesh model and image integration
5. RESULTS AND ANALYSIS
5.1 Mesh Model and Geometric Mapping Analysis
In this research, a third party software, 3D Reshaper 7.1 was
used to texture Bukit Batu Pahat shrine with external digital
images. Nevertheless, integrating both laser scanning and
digital images data was not a straightforward process. The 3D
point cloud model must be transformed into a 3D mesh model.
A mesh model is important for controlling the quality of the
3D model where it keeps only the most relevant points of the
object and optimized the size of the 3D model. Indeed, by
transforming millions of points into one entity, allows other
3D modelling software to make used the model for other
applications such as reverse engineering, animations and
photo-realistic rendering.
The 3D mesh model of Bukit Batu Pahat shrine was quite
impressive although there were several small holes. However,
after enhancing the 3D mesh model using tools like smoothing
and filling holes, the 3D mesh model was ready to be textured
with high-resolution digital images. Besides that, there was no
spike or noise in the 3D mesh model because the points have
been clean-up before meshing process started. A geometric
mapping accuracy was applied onto the 3D mesh model
where, several measurements of the shrine were made on site
using measurement tape and compare with the measurements
done virtually in 3D Reshaper 7.1. Fig-12 illustrates the
measurement conducted in 3D Reshaper 7.1 while the
measurements results are tabulated in Table-4.
Based on the measurement results tabulated in Table-4, the
differences between two measurement methods was 1
millimetre (minimum) to 7 millimetres (maximum). On the
other hand, the RMS value for the measurements is
approximately 4 millimetres. By referring to Table-2, the
geometric accuracy requirement for detailed monument
recording was between ± 2 millimetres until ±5 millimetres.
Thus, from the result obtained, it showed that laser scanning
data was capable of providing the required geometric accuracy
in detailed monument recording.
Fig-12: Measurements of the shrine for geometric accuracy
validation

__________________________________________________________________________________________
Table-4: Comparison between on site measurements and
virtual measurements
Segment
On-Site
Measurement
(m)
Virtual
Measurement
(m)
Deviation
(m)
a 10.596 10.603 0.007
b 23.966 23.963 0.003
c 10.320 10.314 0.006
d 13.462 13.458 0.004
e 10.654 10.656 0.002
f 10.730 10.733 0.003
g 0.374 0.373 0.001
Root mean square (RMS) value 0.004
5.2 Photo-Realistic Model and Feature Mapping
Analysis
Although the geometric accuracy of the mesh model was
impressive and the model has been used as a base for texturing
process, it does not mean that the 3D photo-realistic model
would turn out to be as good as the mesh model result. Hence,
an analysis called feature mapping analysis was carried out to
evaluate the quality of the photo-realistic model by comparing
the 3D photo-realistic model with digital images of the shrine.
A feature mapping accuracy was chosen as an approach to
evaluate how closed was the feature in 3D photo-realistic
model with the real monument. Several identical features of
the 3D photo-realistic model were compared with the
corresponding images. The identical features such as the
colour, holes and broken rock were chosen as the criteria to
evaluate the quality of the 3D photo-realistic model.
Thus, for the first parameter in feature mapping analysis, the
colour of the 3D photo-realistic model was compared with
high resolution images to know the accuracy of merging two
different types of data. Several areas were selected for each
parameter to be analyzed and the results are tabulated in
Table-5, Table-6 and Table-7 respectively. Each area
contained one to three selected spots that were used for
detailed comparison.
Table -5: Feature mapping analysis – comparing colours of
the shrine
Feature Mapping Analysis (Colour)
3D Photo-Realistic
Model
High Resolution
Images
Simila
r Spots
3/3
3/3
3/3
2/3
Quality (%)
[Number of similar spot/Total spot] x 100
91.67
By referring to Table-5, the colour quality of Bukit Batu Pahat
shrine was approximately 91.67% matched with the real
objects. The spots chosen were based on the uncommon
colour or noticeable features appeared in the area such as a
pink target marked on the shrine, white colour at certain rock
and peculiar rock location. Eleven out of twelve spots
possessed the same colour like the real monument. Only one
spot was detected to be difference with the image. This might
be due to the location of the spot, which was at the edge of the
image. Thus, the image was slightly distorted and affected the
colour of the rectangular rock (covered with green moss).
Consequently, the rock became non-rectangular shape because
the colour was only applied at certain part of the rock.

__________________________________________________________________________________________
Table-6: Feature mapping analysis – comparing holes of the
shrine
Feature Mapping Analysis (Holes)
3D Photo-Realistic
Model
High Resolution
Images
Similari
ty
3/3
1/2
Quality (%)
[Number of similar spots/Total spot] x 100 80.00
Bukit Batu Pahat shrine is a structure built from rocks
obtained from a river. The structure is not cemented to cover
holes between the rocks. As a result, there are many holes in
the structure. These holes are considered as one parameter to
analyze the quality of the photo-realistic model. Based on
Table-6, 80% of holes are well presented in the 3D photo-
realistic model. The shape of the holes is maintained without
any distortion. However, there was one hole, which was not
well presented in the model. This was caused by the shape of
the mesh model that was not quite good. Thus, when an image
tried overlapped with the mesh mode, the image tended to
cover the whole mesh even the structure was a little bit
different than the real one. In conclusion, features like holes
could be well-preserved in the photo-realistic model if the
mesh model or the base model was constructed accurately.
Table-7: Feature mapping analysis – comparing broken rock
of the shrine
Feature Mapping Analysis (Broken Rock)
3D Photo-Realistic
Model
High Resolution
Images
Similari
ty
1/1
2/2
Quality (%)
[Number of similar spots/Total spot] x 100 100.00
Broken rocks were the third parameter used to analyze the
quality of the 3D photo-realistic model. This feature was
selected because it was the most obvious feature of the shrine.
Yet the existing of broken structure can help archaeologist to
identify the changes or deformation of the monument and they
can plan for maintenance process. According to Table-7, the
broken rocks are 100% textured without changing their shape.
However, there is a small broken part undergoes image
distortion because the image is not attached perpendicular to
the broken part. This problem can be overcome by capturing
an image perpendicular with the broken surface and map it to
the mesh model.
Generating a 3D photo-realistic model of a structure is not an
easy task. It requires proper data collection planning,
systematic data management and smart way of data
processing. In this research, the structure of the shrine is quite
big thus required longer time to integrate both image and
scanner data. An automatic texture menu in 3D Reshaper 7.1
is very helpful to achieve the main objective of this research.
Several problems were realized during integrating the data.
First of all, the image must be captured perpendicular and
closer to the structure (fit with the camera frame). This is to
ensure that the process of texturing run smoothly. Besides
that, it is important to capture the image with common features
such as the edge to ease the process of marking the
corresponding points.
CONCLUSIONS
This research was conducted to produce 3D photo-realistic
model with required geometric accuracy of archaeological
monument via multi-sensors data integration. In conclusion,
the resulting 3D photo-realistic model was convincing
because, most of the important features like the colour or the
texture, existing holes and broken parts were well preserved in
the virtual model. Hence, archaeologists and researchers can
use the model to do their research or analysis without
travelling to Bujang Valley. The result of the 3D photo-
realistic model can be visualized in Reshaper Viewer 2013
where user can display the model (zoom and rotating) and do
some measurement onto the model.

__________________________________________________________________________________________
ACKNOWLEDGEMENTS
We would like to express our gratitude and thanks to the
Director and staff of Bujang Valley Archaeological Museum,
Malaysia for their supportive cooperation, priceless help and
guidance throughout the data collection and interviewing
session.
REFERENCES
[1]. Guarnieri, F. Marton, A. Vettore. (1999). 3D Modelling of
Real Artistic Objects with Limited Computers Resources.
Centre of Cartography and Photogrammetry (CIRGEO),
University of Padua, Italy. Retrieved on 06 January 2012 from
cipa.icomos.org/fileadmin/papers/olinda/99c415.pdf
[2]. Aguilera, D.G. and Lahoz J. G. (2006). Laser scanning or
Image-based Modelling? A Comparative through the
Modelization of San Nicolas Church. International Archives
of Photogrammetry and Remote Sensing, Volume XXXVI,
B5, Dresden
[3]. Campana, S., Sordini, M. and Rizzi, A. (2009). 3D
Modelling of Romanesque Church in Tuscany: Archaeological
Aims and Geomatics Techniques. In proceedings of 3rd
International Workshop 3D ARCH 2009, Trento, Italy: 25-28
[4]. Remondino, F. (2010). Sensors and Data Integration for
3D Modelling Applications. International Summer School “3D
Modelling in Archaeology and Cultural Heritage 2010”, 17-20
June 2010, Durham, United Kingdom
[5]. Lambers, K. and Remondino, F. (2007). Optical 3D
Measurement Techniques in Archaeology: Recent
Developments and Applications. First publication in: Layers
of Perception: Proceedings of the 35th International
Conference on Computer Applications and Quantitative
Methods in Archaeology (CAA), Berlin, Germany, April 2-6,
2007 / Ed. by Axel Posluschny ... Bonn: Habelt, 2008, pp. 27-
35
[6]. Campana, S., Sordini, M., and Remondino, F., (2008)
Integration of Geomatic Techniques for the Digital
Documentation of Heritage Areas. EARSEL workshop
Advances in Remote Sensing in the Archaeology and the
management of Cultural Heritage, Rome, Italy
[7]. Campana, S. and Remondino, F. (2009). Fast and Detailed
Digital Documentation of Archaeological Excavations and
Heritage Artifacts. 35th CAA Conference (Computer
Applications and Quantitative Methods in Archaeology), pp
36-42, Berlin, Germany, April 2007.
[8]. Heine, E. (2004). 3D Risk Mapping: Theory and Practice
on Terrestrial Laser Scanning Training Material Based on
Practical Applications. (4th
ed.). VLAAMS Leonardo da Vinci
Agentschap: Netherland
[9]. Georgopoulos, A. and Ionnidis, C. (2004).
Photogrammetric and Surveying Methods for the Geometric
Recording of Archaeological Monuments. Workshop -
Archaeological Surveys. FIG Working Week 2004. 22-27
May. Athens, Greece
[10]. Fang, S., George, B. and Palakal, M. (2009). Automatic
Surface Scanning of 3D Artifacts. The International Journal
of Virtual Reality. 8(4), pg. 67-72
[11]. Guidi, G., Remondino, F., Russo, M., Menna, F. and
Rizzi, A. (2008). 3D Modeling of Large and Complex Site
Using Multi-sensor Integration and Multi-resolution Data. The
9th International Symposium on Virtual Reality, Archaeology
and Cultural Heritage VAST (2008), 2-5 December. Braga,
Portugal, 85-92
[12]. Abdullah Mah (2005). Quickbird Versus Aerial Photos
in Identifying Man-Made Objects. 1st National GIS
Symposium in Saudi Arabia. 28 July, Jeddah, Saudi Arabia
[13]. Gonzo, L., Voltolini, F., Girardi, S., Rizzi, A.,
Remondino, F. and El-Hakim, S. F. (2007). Multiple
Techniques Approach to the 3D Virtual Reconstruction of
Cultural Heritage. Eurographics Italian Chapter Conference
2007
[14]. Kadobayashi, R., Kochi, N., Otani, H. and Furukawa, R.
(2004). Comparison and Evaluation of Laser Scanning and
Photogrammetry and Their Combined Use for Digital
Recording of Cultural Heritage, ISPRS Comm. V Sym.
[15]. Ong, C. W., Cheong, S. W., Zulkepli, M. and Halim, S.
(2010). 3D Documentation And Preservation of Historical
Monument using Terrestrial Laser Scanning. Geoinformation
Science Journal, Vol. 10, No. 1, 2010, pp: 73-90
[16]. News Straits Times (2008). Zoom Malaysia – Gone but
not Forgotten. (2008, December 10). New Straits Times:
Tourism Special, pg. 1-11
[17]. Nik Hassan Shuhaimi Nik Abdul Rahman (2004). The
Encyclopedia of Malaysia: Early History. (4th ed.). Kuala
Lumpur, Malaysia: Archipelago Press
[18]. Nik Hassan Shuhaimi Nik Abdul Rahman (2006).
Kepentingan Arkeologi Dalam Kerja Konservasi Monumen
Dan Tapak Tanah Bersejarah. Course on Monument
Conservation and Archaeological Sites. Pulau Pinang
[19]. Patel, A. (2009). High Definition Documentation to
Enhance the Interpretation of Archaeological Structures and
Sites. Degrees of Master in Architecture, Texas Tech
University, USA
[20]. White, M., Mourkoussis, N., Darcy, J., Petridis, P.,
Liarokapis, F., Lister, P., Walczak, K., Wojciechowski, R.,
Cellary, W., Chmielewski, J., Stawniak, M., Wiza, W., Patel,
M., Stevenson, J., Manley, J., Giorgini, F., Sayd, P. and
Gaspard, F. (2004). Proceedings of the Computer Graphics
International 2004. 16 – 19 June 2004. Greece, pg 1 – 4.
[21]. Alshawabkeh, Y. and Haala, N. (2004). Integration of
Digital Photogrammetry and Laser Scanning for Heritage
Documentation. International Archives of Photogrammetry
and Remote Sensing. Vol.5. Istanbul, Turkey
[22]. Al-Hanbali, N. N., Albayari, O., Saleh, B., Almasri, H.
and Baltsavias, E. (2006). Macro to Micro Archeological
Documentation: Building a 3d GIS Model for Jerash City and
Artemis Temple. Alias Abdul-Rahman, Zlatanova, S. and
Coors, V. (Eds.) Innovation in 3D Geo Information Systems.
New York: Springer Berlin Heidelberg

__________________________________________________________________________________________
BIOGRAPHIES
Nurul Shahida Binti Sulaiman
Researcher in Photogrammetry & Laser
Scanning Research Group (PLSRG) Bsc in
Geomatic Engineering, Faculty of
Geoinformation & Real Estate, Universiti
Teknologi Malaysia Area of interest:
LiDAR and 3D modelling
Erna Warnita Binti Bachad Researcher
in Photogrammetry & Laser Scanning
Research Group (PLSRG) Bsc in
Geoinformation Science, Faculty of
Geoinformation & Real Estate, Universiti
Teknologi Malaysia Area of interest: GIS
and LiDAR
Dr Albert Kon-Fook Chong Senior
Lecturer School of Civil Engineering and
Surveying, University of Southern
Queensland Phd from Washington Area of
interest: Photogrammetry, 3D modelling,
Information Database Design
Assoc. Professor Dr Zulkepli Bin Majid
Associate Professor Faculty of Geo
information & Real Estate, Universiti
Teknologi Malaysia Phd from Malaysia
Area of interest: Photogrammetry, 3D
modelling, Laser Scanning, GIS
Professor Halim Bin Setan Professor
Faculty of Geoinformation & Real Estate,
Universiti Teknologi Malaysia Phd from
United Kingdom Area of interest: Laser
Scanning, Deformation Survey and
Surveying Engineering

Generating three dimensional photo-realistic model of archaeological monument via multi-sensors’ data integration

More Related Content

What's hot (17)

Similar to Generating three dimensional photo-realistic model of archaeological monument via multi-sensors’ data integration (20)

More from eSAT Journals (20)

Recently uploaded (20)

Generating three dimensional photo-realistic model of archaeological monument via multi-sensors’ data integration