Content uploaded by Elisavet Telioni
Author content
All content in this area was uploaded by Elisavet Telioni on Mar 20, 2023
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=ysre20
Download by: [National Technial University of Athens], [Doctor E. C. Telioni] Date: 01 March 2016, At: 01:03
Survey Review
ISSN: 0039-6265 (Print) 1752-2706 (Online) Journal homepage: http://www.tandfonline.com/loi/ysre20
The contribution of laser scanning technology
in the estimation of ancient Greek monuments'
deformations
G. D. Georgopoulos, E. C. Telioni & A. Tsontzou
To cite this article: G. D. Georgopoulos, E. C. Telioni & A. Tsontzou (2016): The contribution
of laser scanning technology in the estimation of ancient Greek monuments' deformations,
Survey Review
To link to this article: http://dx.doi.org/10.1179/1752270615Y.0000000035
Published online: 29 Feb 2016.
Submit your article to this journal
View related articles
The contribution of laser scanning technology
in the estimation of ancient Greek monuments’
deformations
G. D. Georgopoulos
1
, E. C. Telioni*
1
and A. Tsontzou
2
Terrestrial laser scanning technology is nowadays more and more used for the documentation of
cultural heritage monuments. The thorough exploitation of the main advantage of terrestrial laser
scanners (TLS) that is the acquisition of extremely dense discrete points in a relatively small time
period leads to detailed 3D representation of the monument, overcoming possible difficulties
such as limited accessibility. Most often, this 3D representation is used for the monument’s
documentation as well as for virtual tours in, out or around it. This detailed documentation can be
used for one more purpose: the estimation of the deformation that its elements have experienced
through the centuries and, following, the monument’s restoration, provided that, its’ initial,
constructional geometry is known. In this paper, research towards this goal is presented, dealing
with the estimation of the deformations of a column of the ancient temple of Zeus in Nemea Greece
using TLS technology. This column is standing erect since the temple’s construction in 330 BC
and it has been subject to serious deformations because of various causes.
Keywords: Cultural heritage monuments, Ancient Greek monuments, Laser scanning technology, Estimation of deformations, Geodetic methodology
Introduction
In recent years, an increasing interest on the application
of terrestrial laser scanners (TLS) in deformation
estimation is observed (Alba et al., 2006; Kersten et al.,
2009; Roberts and Hirst, 2005). This interest arrises
from the advantages of TLS that can be summarised to
the direct 3D coordinates estimation of a dense points
set on the object monitored, the high degree of
automation, the user-friendly software that accompanies
the hardware and the creation of the examined object’s
model (Monserat et al., 2007). An important and very
interesting part of deformation monitoring using TLS is
that concerning cultural heritage monuments
(Castagnietti et al., 2012; Pesci et al., 2011; Teza and
Pesci, 2013; Pieraccini et al., 2014). This monitoring
deals not only with ‘current time’ deformations because
of various causes (seismic hazard, underground works,
etc.) but also with the deformations the parts of the
monument have undergone through the centuries: this is
of crucial importance when planning interventions con-
cerning restoration of the monument.
The possibility of obtaining a large amount of object
points is considered as an extremely important advan-
tage of laser scanners compared to classic geodetic
methodology. In the later case, the estimation of the
object’s deformation is achieved through the
establishment of a geodetic control network and the
monitoring of discrete control points on the monument
under consideration. On the contrary, when TLS
technology is applied, the object’s deformations are
determined after the creation of a 3D model from the
points cloud either through an appropriate software or
through the estimation of a surface that fits best, using
least squares techniques (Gruen and Akca, 2005;
Pesci et al., 2011; Teza and Pesci, 2013). From this
modelled surface, all the information, concerning the
way the monument’s elements have deformed, is derived.
This is achieved through the comparison of the modelled
surface with the initially constructed one. The high
redundancy of points ameliorates the accuracy in
deformation estimation, provided that the appropriate
methodology is applied, in order to estimate statistically
significant small scale deformations. The later is
extremely important in cases of restoration of cultural
heritage monuments where high accuracy is crucial for
correct interventions. This fact holds especially for the
restoration of monuments of the Greek Antiquity
(temples, theatres, etc.), since it is known that they were
constructed on the basis of strict plans, demanding high
accuracy in construction, and therefore, the same accu-
racy must be achieved during the restoration project.
The purpose of this paper is to present an approach
on the estimation of ancient Greek monuments’ defor-
mations that takes advantage of laser scanner technol-
ogy but is based on the principles of classic geodetic
methodology. This approach is applied – and presented
here – for the estimation of the deformations of a
column of the ancient temple of Zeus in Nemea, Greece,
that is standing erect since the temple’s construction.
1
School of Rural and Surveying Engineering, National Technical
University, 9, H. Polytechneiou Str., Zografou, Athens 157 80, Greece
2
18, Ioannou Theodorou Str., Koropi, Athens 19400, Greece
*Corresponding author, email telioni@survey.ntua.gr
Ñ2016 Survey Review Ltd
Received 08 December 2014; accepted 04 May 2015
DOI 10.1179/1752270615Y.0000000035 Survey Review 2016 1
Downloaded by [National Technial University of Athens], [Doctor E. C. Telioni] at 01:03 01 March 2016
The proposed methodology
In the majority of cases, deformations in ancient Greek
monuments are determined from direct measurements.
These measurements are of high metric accuracy but are
time consuming, may demand special construction
(scaffold) for the measurements, and, what is more
important, cannot be related to each other.
In the cases where classic geodetic methodology is
used, the deformations are estimated using a geodetic
control network. This network consists of reference
points, established in stable areas in the vicinity of the
monument, and control points, established in carefully
selected positions on the monument itself. The elements
of the network are measured and adjusted using least
squares and the coordinates of the network’s points are
estimated in a local or global reference frame.
On the contrary, when using TLS technology,
the monument’s deformations are estimated from the
entire modelled object rather than from discrete points
of the cloud, which are difficult to be identified and
isolated. The points of a local geodetic network, esta-
blished, measured and adjusted for this reason in the
area of the monument, are used as scan positions.
In order to ensure the accurate registration of successive
scans as well as the high accuracy of the geo-referenced
point clouds, special (planar or spherical) targets are
positioned at the network’s points. Finally, the extracted
deformations are referred to the same reference frame of
the network and can be associated and compared
to each other, so that a complete and detailed ‘picture’ of
the deformed monument is derived.
As far as the scanning resolution is concerned, it is
proposed to perform the scans with maximum sampling
provided by the scanner. Before any operation, it is
necessary to calibrate both its hardware and software,
in order to ensure the proper operation of the scanner.
The procedure of the monument’s modelling requires
the creation of a surface best-fitting to the points cloud.
The creation and elaboration of the optimum surface are
a complex process, mainly because of the fact that it is
non-parametric and the primary data are often quite
noisy. The basic steps of the creation of the surface
include:
- The data pre-processing for noise reduction and
deletion of those points that do not belong to the
monument.
- The specification of the object’s features that should be
maintained during the modelling (e.g. edges or break
lines). For example, when modelling the surface of an
ancient temple’s column, the break lines, indicating the
parts of the column (drums, capital, etc.) must be
maintained.
- The creation of triangular or tetrahedral meshes and
the generation of the surface. The generated surface
can also be mathematical (parametric) if the geometry
of the object is simple. The majority of the used
software performs the technique of triangulation,
where a set of points is changed to a polygonal mesh.
Further refinement of the generated surface, such as
hole filling and smoothening, is usually required.
- Right after the surface generation, specifically selected
point sets are extracted. These points belong to planes,
lines, curves, etc. that describe the object’s current
geometry. From these extracted point sets, the best-
fitting plane or line is determined, using least squares
techniques, and is compared to the initial,
constructional one in order to estimate statistically
significant deformations of the monument’s elements.
It is very important to have a good knowledge of the
monument’s initial, constructional geometry for the
correct interpretation of the results. For example,
when dealing with the estimation of ancient temples’
deformations, one must keep in mind that the main
lines expressing the monument’s geometry are
either curved or inclined with rates in correlation to
each other.
Deformation estimation of a column of
the ancient temple of Zeus in Nemea,
Greece
The proposed methodology was applied for the
estimation of the deformations of a column of the
ancient temple of Zeus in Nemea. The temple, of Doric
order, was constructed in 330 BC. It was an important
Greek sanctuary and was the centre of the Panhellenic
Nemean Games. The temple had an exterior colonnade
(called peristyle) with six columns on the short sides and
12 on the long ones. The systematic demolition of the
temple began between 600 and 700 AD, in order to use
its elements for the construction of a church situated
about 100 m south of the temple. The external Doric
columns were knocked down and about 1300 parts of
the temple were removed. Only three of the 36 Doric
columns remained in place, one of which is the column
under consideration. The parts of the temple remaining
in place (the stepped base and the three Doric columns)
show significant displacements because of human
destruction (mostly in the early Christian period) as well
as because of settlements, since the foundation of the
temple is not on solid mass. The restoration works on
the temple of Zeus began in 1984. Up till now, two
restoration programmes have been completed, during
which, part of the temple’s stepped base (the crepis) and
six of the peristyle columns – those of the northeastern
corner – were completely restored (summer 2013).
The column under consideration is the only one of the
peristyle columns, that stands erect since the temple’s
construction (Fig. 1). It is standing next to the lately
restored columns at the eastern side of the temple.
The column has a total height of 9.726 m (without the
capital), and consists of 13 superimposed parts (drums)
that have a mean height of 0.75 m. Each drum has
20 carved shafts ( flutes) and was connected to the
adjacent drums through a wooden dowel. The drums’
diameters reduce gradually from 1.522 m (lower face of
first drum) to 1.237 m (upper face of thirteenth drum).
A typical cross-section of a drum is depicted in Fig. 3.
Significant displacements and rotations between the
column’s drums are visible. In order to plan restoration
interventions, it was decided to estimate the column’s
deformations using TLS technology.
Data acquisition – modelled surface
A geodetic control network was already established at
the temple’s surroundings, as well as on its body, for the
needs of the two restoration programmes mentioned
above (Georgopoulos and Telioni, 2005). Three new
points were established for the scanning purposes.
The network’s elements were measured and adjusted,
Georgopoulos et al. Contribution of laser scanning technology
2Survey Review 2016
Downloaded by [National Technial University of Athens], [Doctor E. C. Telioni] at 01:03 01 March 2016
and the point’s coordinates in a local reference system
were estimated with an accuracy of few millimetres
(Tsontzou, 2014).
Terrestrial laser scanner ScanStation2 of Leica Geo-
systems was used for the column’s scanning. For the
referencing and registration of the points clouds, HDS
sphere targets were used. The instrument has, according
to the manufacturers, an accuracy of +4 mm in distance
measurement, an angular accuracy of +60 mrad (120),
while the accuracy in point positioning is +6 mm. Taking
into account the above mentioned accuracies, as well as
the error because of the beam’s diameter, the centring and
levelling errors of the scanner and targets (Lichti and
Gordon, 2004), the horizontal and vertical accuracy of
direct geo-referencing was estimated to be +5 mm.
Four of the network’s points were used for the
scanning. The Cyclone software was used for the
definition of the scan parameters and the processing of
the field data. The scanning resolution was set to 2 mm.
When each scan was completed, the point cloud was
georeferenced directly to the local reference frame of the
network. The geo-referencing errors were small, up to
4 mm. The whole procedure was completed in a 5-h
interval. The modelled surface of the column is depicted
in Fig. 2 (Tsontzou, 2014).
Estimation of deformations
After the creation of the column’s modelled surface
(Fig. 2), the 13 drums of the column were isolated
from each other. For each drum, three horizontal cross-
sections were created: at the upper and lower face of the
drum and one at the middle, 39 sections as a whole. The
exact position of each section depended on the existing
damages of the drum. These 39 sections were extracted
in Cad environment and converted into point sets.
In Fig. 3, the modelled surface of a drum is depicted,
together with the result of the extracted middle section in
Cad environment and the corresponding point set.
For each point set, after it was extracted in Cad
environment, the coordinates of the inner points of the
flutes were determined. These points (one for each flute,
20 as a whole) lie on the circumference of the circle
escribed in the section (Fig. 3). The circle’s components
(centre’s coordinates and radius) were estimated using
the following equation (1)
ðxi2x0Þ2þðyi2y0Þ22R2¼0ð1Þ
where
x
0
,y
0
: the coordinates of the escribed circle’s
centre
x
i
,y
i
: the coordinates of the inner point of i flute
(i¼1,..., 20)
R: the escribed circle’s radius
Thus, 20 observation equations were formulated for
each section. After the adjustment, the coordinates
(x
0
,y
0
) of the circle’s centre, as well as its radius R, were
estimated. Their standard deviations ranged between
¡0.1 mm and ¡1.3 mm.
In order to check the precision achieved, the diameters,
corresponding to the section at the lower face of each of the
1 Column K31 of the temple of Zeus in Nemea
2 Modelled surface of the column
Georgopoulos et al. Contribution of laser scanning technology
Survey Review 2016 3
Downloaded by [National Technial University of Athens], [Doctor E. C. Telioni] at 01:03 01 March 2016
13 drums, were compared to the direct measurements,
made with a mechanical calliper that had an accuracy of
¡0.1 mm (Zambas, 2000). The diameters were, also,
compared to the corresponding ones of the adjacent
column that, as it has been mentioned above, was restored
recently. In both cases, differences did not exceed 2 mm.
The constructional, initial positions of the centres of
the columns’ bases (lower face of the first drum) are
determined from previous surveys of the temple, for the
restoration projects (Georgopoulos and Telioni, 2005).
They belong to a straight line, parallel, at a distance of
0.900 m, to the upper step (the stylobate) of the temple’s
base. The distance between the centres is constant
(3.745 m), with the exception of the four corner
columns, where the distance is about 0.20 m smaller.
Thus, the initial, constructional position of the column
monitored is known. It was realised, however, from the
estimated coordinates of the centre of the column’s
base that the column has moved, as a rigid body,
from its initial, constructional position 0.066 m with a
south-western direction.
As it has been mentioned above, for the correct
interpretation of the estimated deformations, a good
knowledge of the monument’s constructional geometry
is needed. Thus, when dealing with deformations of
ancient columns, it must be kept in mind that the peristyle
columns were constructed with a given inwards
inclination with respect to the plumb line. This inclination
was imposed, because the surface of the temple’s stepped
base (on which the columns were erected) was constructed
curved, with different transversal and longitudinal
gradients. The inwards inclination of the columns’ axis
has a rate equal to the transverse inclination of the base
and can be estimated from direct measurements on the
columns’ first drum (Zambas, 1998). For the temple of
Nemean Zeus, this inwards inclination, of all the columns’
axes, is estimated (from direct measurements with a
mechanical calliper) 3.5‰ (Zambas, 2000). Therefore, the
initial, constructional positions of the drums can also be
estimated, and compared with the current ones, in order
to estimate correctly the drums’ displacement vectors.
The displacement vectors between every two
successive drums were estimated and their statistical
significance was tested for confidence level 95%. For this
purpose, the centres of each drum’s middle section were
used. Significant displacements ranging up to 21 mm
were detected (Tsontzou, 2014). The displacement
vectors (length d
ij
and direction a
ij
) are depicted in
Table 1 and presented in Fig. 4.
From Fig. 4, it can be seen that, besides the 0.066 m
column’s movement as a rigid body, its eight first drums
show a movement, varying in magnitude, but with the
same north–northwestern direction. For the rest five
drums, the direction of the displacement vectors changes
completely, having a direction towards the south.
In Fig. 4, the initial, constructional position of the
column’s drums centres is also given for comparison
reasons.
3 Cross-section of a drum and the corresponding point set. The circle described in the section is also presented
Table 1 Displacement vectors between the column’s
successive drums
Drums d
ij
(mm) s
d
(mm) a
ij
(g)
First–second 12 ¡0.6 325
Second–third 4 ¡0.6 367
Third–fourth 4 ¡0.5 354
Fourth–fifth 12 ¡0.7 347
Fifth–sixth 6 ¡0.8 351
Sixth–seventh 6 ¡0.7 346
Seventh–eighth 21 ¡0.6 365
Eighth–ninth 9 ¡0.7 344
Ninth–tenth 3 ¡0.8 267
Tenth–eleventh 8 ¡0.8 201
Eleventh–twelfth 5 ¡0.7 175
Twelfth–thirteenth 5 ¡0.8 173
4 Displacement vectors between the column’s successive
drums
Georgopoulos et al. Contribution of laser scanning technology
4Survey Review 2016
Downloaded by [National Technial University of Athens], [Doctor E. C. Telioni] at 01:03 01 March 2016
The displacement vectors of all the drums, with
respect to the first one, were also estimated (Table 2) and
are presented in Fig. 5. Displacements ranging up to
7 cm were detected.
Apart from the drums’ linear displacements, the
rotation Qof each drum (with respect to the first one),
and the corresponding arc length on the escribed circle,
were also estimated (Table 3, Fig. 6).
As it can be seen, all the drums show a counter-
clockwise rotation with respect to the first one: the
maximum rotation of 5 g occurs from the eighth up to
the eleventh drum that seem to have rotated as a
rigid body.
Conclusions
The proposed methodology and its application show
that, laser scanning technology can be an indispensable
tool for the estimation of the deformations that their
parts have undergone during the centuries of their life.
The establishment of a high precision geodetic network
and the use of its points as stations for the monument’s
scanning ensures the prerequisite of accurate registration
and geo-referencing of points clouds. In fact,
geo-referencing errors did not exceed +3 mm in the case
study presented in the context of this paper.
The main difference on deformation estimation when
using TLS technology is that one deals with big point
sets rather than discrete points as in classic geodetic
monitoring. To overcome this problem, the approxi-
mation of carefully selected point sets with the appro-
priate line or surface, using least squares techniques, is
proposed in order to determine the structural geometry
of the object examined. In the case examined above, this
led to the estimation of the necessary geometric features
with +1 mm accuracy. It is important to note the good
agreement of these features from the directly measured
ones, the differences been all statistically insignificant.
Table 2 Displacements of the column’s drums with respect
to the first one
Drums d
ij
(mm) a
ij
(g)
First–second 12 325
First–third 15 334
First–fourth 18 337
First–fifth 30 341
First–sixth 35 343
First–seventh 41 343
First–eighth 61 351
First–ninth 70 350
First–tenth 71 347
First–eleventh 66 341
First–twelfth 62 338
First–thirteenth 58 336
5 Displacements of the column’s drums as seen from the
southern and eastern side of the temple
Table 3 Rotations of the column’s drums with respect to the
first one
Drums Rotation Q(g) Arc length (mm)
First–second 21.1 13
First–third 20.4 5
First–fourth 20.6 7
First–fifth 21.9 22
First–sixth 21.6 18
First–seventh 22.0 23
First–eighth 24.5 50
First–ninth 24.7 51
First–tenth 25.0 54
First–eleventh 25.0 52
First–twelfth 22.0 21
First–thirteenth 21.8 18
6 Rotations of the column’s drums with respect to the first
one and the corresponding linear movements
Georgopoulos et al. Contribution of laser scanning technology
Survey Review 2016 5
Downloaded by [National Technial University of Athens], [Doctor E. C. Telioni] at 01:03 01 March 2016
The estimated deformation parameters (displacement
vectors and drums’ rotations) had the same accuracy of
+1 mm. The selection of the appropriate line or surface
for the approximation of the extracted point sets is of
utmost importance, therefore, the cooperation with
architects specialised in restoration is a prerequisite.
Moreover, a good knowledge of the monument’s initial
geometry is needed, for the correct interpretation of the
results of deformation estimation.
It must be pointed out that, since all the points clouds are
referenced to the network’s reference frame, the estimated
deformations are related to each other, thus giving a
complete ‘picture’ of the way the monument has deformed.
This can be used for the estimation of the monument’s
vulnerability and the decision of needed interventions.
Taking into account the small time interval needed for
the points cloud acquisition, when compared to the
amount of information gained, the proposed method-
ology can be applied for a quick, but accurate, check of
the monument’s condition, after a sudden event, such as
an earthquake, has occurred.
References
Alba, M., Guissani, A., Roncoroni, F., Scaioni, M. and Valgoi, P. 2006.
Geometric modeling of a large dam by terrestrial laser scanning.In:
XXIII FIG Congress, Shaping the change, Munich, Germany,
October 8–13.
Castagnietti, C., Bertacchini, E., Capra, A. and Dubini, M. 2012.
Terrestrial laser scanning for preserving cultural heritage: analysis of
geometric anomalies for ancient structures. In: FIG Working Week
2012. Knowing to manage the territory, protect the environment,
evaluate the cultural heritage, Rome, Italy, 6–10 May.
Georgopoulos, G. and Telioni, E. 2005. Geodetic surveys at the temple
of Zeus, in Nemea, Greece, Technical report (in Greek).
Gruen, A. and Akca, D. 2005. Least squares 3D surface and curve
matching. ISPRS Journal of Photogrammetry and Remote
Sensing, 59(3), pp.151–74.
Kersten, T., Sternberg, H. and Mechelke, C. 2009. Geometrical building
inspection by terrestrial laser scanning. In: FIG Working Week
2009. Surveyors key role in accelerated development, Eilat, Israel,
3–8 May, 2009.
Lichti, D. and Gordon, S. 2004. Error propagation in directly
georeferenced terrestrial laser scanner point clouds for cultural
heritage recording. In: FIG Working Week 2004, Athens,
Greece, 22–27 May, 2004.
Monserrat, O. and Crossetto, M. 2007. Deformation measurement
using terrestrial laser scanning data and least squares 3D
surface matching. ISPRS Journal of Photogrammetry and
Remote Sensing, 63, pp.142–54.
Pesci, A., Casula, G. and Boschi, E. 2011. Laser scanning the Garisenda
and Assinelli towers in Bologna (Italy): detailed deformation
patterns of two ancient leaning buildings. Journal of Cultural
Heritage, 12, pp.117–27.
Pieraccini, M., Dei, D., Betti, M., Bartoli, G., Tucci, G. and Guardini,
N. 2014. Dynamic identification of historic masonry towers
through an expeditious and no-contact approach: application to
the ‘‘Torre del Mangia’’ in Siena (Italy). Journal of Cultural
Heritage, 15, pp.275–82.
Roberts, G. and Hirst, L. 2005. Deformation monitoring and analysis of
structures using laser scanners. In: FIG Working Week 2005 and
GSDI-8. From Pharaohs to Geoinformatics, Cairo, Egypt, 16–21
April, 2005.
Teza, G. and Pesci, A. 2013. Geometric characterization of a cylinder
shaped structure from laser scanner data: development of an
analysis tool and its use on a leaning bell tower. Journal of
Cultural Heritage, 14, pp.411–23.
Tsontzou, A. 2014. Deformation estimation and study of the optical
refinements of column K31 of the Temple of Zeus, in Nemea,
using terrestrial laser scanner., Diploma thesis (in Greek).
School of Rural & Surveying Engineering, N.T.U. Athens, 180 p.
Zambas, K. 1998. The refinements of the Parthenon columns.
Ph D. Thesis (in Greek). N.T.U. Athens, 254 p.
Zambas, K. 2000. Measurements of K31 refinements for restoration
purposes, Technical report (in Greek).
Georgopoulos et al. Contribution of laser scanning technology
6Survey Review 2016
Downloaded by [National Technial University of Athens], [Doctor E. C. Telioni] at 01:03 01 March 2016
