A

w3wdtwma3eb©

K

1

abtHndmia

o(

1d

Journal of Cultural Heritage 10 (2009) 269–274

Case Study

3D Photogrammetric model of Eflatunpinar monumentat the age of Hittite empire in Anatolia

Güngör Karauguz a,∗, Özsen Corumluoglu b, Ibrahim Kalaycı b, Ibrahim Asri c

a Education Faculty, Selcuk University, 42090 Meram-Konya, Turkeyb Engineering and Architecture Faculty, Selcuk University, Turkey

c Professional Higher School, Hacettepe University, Turkey

Received 6 August 2007; accepted 20 August 2008

bstract

Eflatunpinar monument is located about 25 km at the North of Beysehir (Konya–Turkey). The monument which belongs to Hittite Empire Ageas known since xix century and there are several papers written on this monument in the literature. The monument which is 7.02 m wide and.6 m high was built of large tracit block stones. Latest excavation carried out by Konya’s Archeology Museum in 1996 showed that the monumentas built with some other parts and surrounded by a water channel. These results have pushed the scientist into a discussion – as it has usually beenone intensively several times in the past – on environmental arrangements surrounding the monument. Hittite scripts brought to light implicatehat monument surrounded by trees and prohibited someone to cut these trees off by law. Archaeological studies done on Eflatunpinar monumentere carried out by classical techniques. Therefore, those studies were limited by capabilities of classical techniques and were not able to presentore than plan and some cross-sections. By this study, it is made a modern approach which is alternative to classical technique and using a digital

nd virtual environment provided by computer technology. As a result of this study, all measurements can easily be collected by the help of thisD virtual model of the monument with no need to be on the site at any time they are required. Complete 3D model will also provide a numericalvaluation, interpretation and several analyses on monument together with its environment and surroundings. This 3D virtual model approach willring a novelty into Hittite archeology.

2009 Elsevier Masson SAS. All rights reserved.

ling

tfepwmipn

eywords: Eflatunpinar monument; Hittite; Digital photogrammetry; 3D mode

. Introduction

Historical objects, monuments and monumental groups, aressets that reflect the great value coming from the historicalackground of a country because they often represent the his-ory and memory of the communities where they are placed.eritage must normally be protected by local, national or inter-ational authorities, in order to prevent its deterioration andestruction [1]. Such protection is not always as efficient as

ight be expected. Sometimes, it can never take place. If the case

s this, there should be something to be done beforehand suchs implementation of a comprehensive documentation work by

∗ Corresponding author.E-mail addresses: karauguz@msn.com (G. Karauguz),

corumlu@selcuk.edu.tr (Ö. Corumluoglu), ikalayci@selcuk.edu.trI. Kalaycı), iasri@hacettepe.edu.tr (I. Asri).

cbo

imise

296-2074/$ – see front matter © 2009 Elsevier Masson SAS. All rights reserved.oi:10.1016/j.culher.2008.08.013

he help of any conservation technique even if it is speciallyocused to the historical asset in interest. It can be found in sev-ral documentation and conservation techniques available in theractice [2]. These methods are capable of providing the buildingithin a co-ordinates system [3]. The methods and equipmentainly preferred for the documentation and surveying of build-

ngs are [4]: traditional manual methods, topographic methods,hotogrammetric methods and scanning methods. Such tech-iques are taken in the documentation process as tools for theonservation of heritage monuments. The documentation has toe carried out prior to the building’s destruction, transformationr any intervention.

Related to this study, when digital photogrammtery techniques chosen as a conservation technique, it should be created 3D

odels and scaled orthophotos, plans, etc. The course of time,f the progressive deterioration of materials, the exhaustion ofome structures and the damages caused by successive reforms,nd up leading to the destruction and loss of the monuments. The

2 ultur

3mitmec

asipmepesamstoup

tdpsgtltlcro

2

sm

mpB[smts

oBat

kRawK

tmwisa

obs

lg[t

toaotf

otrCrTl

sp

3

tmOipaigc

70 G. Karauguz et al. / Journal of C

D models, orthophotos and plans can then be used to rebuild theonument as similar as how it was before the destruction, even

f in the cases of that intervention is undertaken subsequentlyo partial collapse of the monument. In such occasions, some

aterials are replaced and the restoration works are run by thengineer with respect to those plans, whose most significantriterion is accuracy.

Today, measurement procedures carried out in excavatedrchaeological sites and for objects found at these excavationites represent a very interesting contest where the potential-ties of new digital technologies of photogrammetry can beroductively expressed. But, some difficulties related to the 3Dodelling from digital images can be encountered when mod-

lling surfaces and complex archaeological structures. Severalroblems can also come out at any time with respect to mod-lling of terrain or of manmade objects. We can even meet withome problems related to the data acquisition and integrationnd to the automatic or semi-automatic data handling, or to newethods for data representation and exploration in general. In

pite of these problems and difficulties, digital photogrammetryechnique has been used successfully and proved its potentialityn providing 3D models and documentation of historical mon-ments and archaeological sites with successive studies in theractice.

Anatolia is one of the very few special and unique places inhe world that have played a cradle role for many civilizationsuring the every stage of human history. In Anatolia, there arelenty of historical assets, monuments, constructions, items ando on trying to stay stand on the ground or just underneath theround and buried by soil. They all have been remained fromhe time of these civilizations. All those historical assets reflectife style and carry traces of their owners, user or contractorso our time from the time when those owners lived in this greatand, Anatolia. One of the great civilization left traces of theirivilization in this land is Hittite community. Several historicalemains from Hittites have been found especially in the regionf Konya which is a great city located at the heart of Anatolia.

. The monument of Eflatunpinar

We are not going to give historical and artistical details widelyince this study is especially going to focus on creation of 3Dodel for the monument.The main aim of this study can be stated as building of 3D

odel of a very important archaeological site called as Eflatun-inar close to Sadıkhacı village and at the 25 km North ofeysehir County within Konya province in the hearth of Turkey

5–15]. It is a historical site remained from the Hittites’ era. Thistudy is also expressed as one of the historical documentationissions executed and directed by Corumluoglu who works for

he geodesy and photogrammetry department of Selcuk Univer-ity.

Tarhuntassa was the Hittities’ capital city located at the South

f Central Anatolia [16]. II. Muwattali carred the capital ofogazköy to Tarhuntassa from today’s Bogazköy due to Kaska’sttacks when he went to state of Kadesh battle and to take logis-ic support during the Kadesh battle against to Egypt. This land

qaiA

al Heritage 10 (2009) 269–274

nown as Lower Land (KUR SAPLITI), Tarhuntassa or Hulaiaiver Land was governed by Kurunta who called himself as“Great King”. Governer of the Lower Land before Kuruntaas Hannutti. Later on, competition between IV. Tuthaliia andurunta was seen in this Land.So, it is assumed that Eflatunpinar monument (or open air

emple) was built at that time of this competition (for date ofonument) [7–9,10,12,15,17–20]. Ruins of the open air temple,hich consisted of a pool area surround by a low level site walls

n front of it (in terms of our opinion it is a worshipping Garthurrounded by a low level wall), were brought to the light after anrchaeological excavation by Archaeologic Museum of Konya.

Some figures were seen as they were carved on the facadef temple monument which was built up as enormous stonelocks. Any manuscripts cannot be found on the monument anduggested that those are only figures [14].

These figures were interpreted as hybrid creatures which wereion-bull men carved on the stones surrounding a one god andoddess couple’s figure with tapering conical cap on their head7,15]. In our opinion, they are probability men whirling aroundhemselves, similar to ceremonies seen in Mevlana rituals.

Due to the uncompleted two stone blocks just at 2 m back ofhe monument, there are several opinions on how should be theriginal status of the monument [9–12]. One of these opinionsssumes that arslantash and Fasillar monuments were locatedn Eflatunpinar monument [9]. It was also argued that thesewo monuments were not representing features that follows theeatures on Eflatunpinar monument [15].

In our opinion, these two monuments do not complete eachther [16]. Eflatunpinar monument is also an especially impor-ant historical asset since it was located at the intersection of theoutes from the gates of Hittites opening to the West (Arzawa).ontrary to that, Fasillar was located at the the intersection of the

outes from the gates of Hittites opening to the South (Lukka).herefore, such monuments are assumed as open sacred places

ocated on these routes [16].It is suggested that around this monument and similar’s

hould be surrounded by decoration trees and also it should berohibited to violate such sacred places [21,22].

. 3D modelling by digital photogrammetry

Digital close-range photogrammetry has a few decade his-ory and it has such a proven fact that it is a useful tool for 3D

easurement and modelling in support of heritage recording.ver the past decade or so, we have witnessed a very signif-

cant growth in activity in this area and indeed in close-rangehotogrammetric measurement in general. A number of trendsccount for the broadening use of this technology. The first trends seen in the increasing availability and suitability consumerrade digital cameras for photogrammetric applications. Theseameras are ubiquitous, they have ever higher resolutions and are

uite suitable for accurate measurement required in architecturalnd archaeological recording, forensic measurement, engineer-ng documentation and in numerous other application domains.

second trend is the enhanced capabilities available in 3D

ultur

mcbiotc

3

mtsrcCdaomdpm

pmhCts

vpcoortibs1Iaficsit

4

oapa

G. Karauguz et al. / Journal of C

odelling and visualization, processes which benefit from moreomprehensive measurement and imagery data. A third trend cane related to developments in photogrammetric data processing,ncluding improved computational models, and the developmentf automated image measurement and photogrammetric orien-ation software systems designed for use with low-cost digitalameras.

.1. Photogrammetric processes

In archaeological photogrammetry, we like to fulfil the goal:etric information describing the monument’s construction sys-

em and its spatial layout, starting from the three-dimensionaltructure when the decision taken ends up with conservation,econstruction and even documentation. Nowadays the resultsan be displayed in three different ways: handmade drawings,AD drawings, 3D visualisations and animations. The methodsevised for documentation have some added disadvantages suchs slowness and laboriousness, in some cases, and high cost inther cases, when acceptable results are expected. Thus, theyay not be operative in some circumstances [23]. Close-range

igital photogrammetry can solve those problems, because itrovides reliable representations of historical structures docu-ented in a fast, precise and inexpensive way.To apply close range photogrammetry (in such a specific

roject), after determining the object which is supposed to beodelled, it is necessary to design control points in a way

omogenously distributed all over the facades of monument.ontrol points must be seen clearly and visibly in images and

heir coordinates (X, Y, Z) are known in a reference coordinateystem most probably measured by the help of a traditional sur-

scbr

Fig. 1. Distribution of the control points

al Heritage 10 (2009) 269–274 271

eying technique. This system is preferably defined as its XYlane perpendicular to horizon and Z-axis towards location ofameras (and perpendicular to XY plane). Geometric positionsf control points must form volume, i.e., they must not lie onnly one plane, otherwise collinearity equations will not obtaineasonable solution for unknowns [24]. After determining con-rol points, images have to be taken from different positions. Its not necessary that all control points be visible in all images,ut in each image at least six control points have to be visible toolve collinearity equations since a full camera calibration with1 parameters is necessary in the case of non-metric camera use.n this process, unknown parameters such as camera orientationngles (Omega, Fi, Kapa) and perspective center coordinates andocal length of each camera will be solved. Then, if a real pointn object space was visible in more than one image, object spaceoordinates of that point can be computed by method of inter-ection. Finally, we have image coordinates of a specific pointn all images, so it is possible to calculate ground coordinates ofhat point by intersection method in a least square manner.

. Case study

In order to experiment the potentiality of image-based meth-ds for the metric documentation and the 3D reconstruction ofn important archaeological site such as the one studied in thisroject, a close-range photogrammetric survey of the monumentnd the surrounding details was conducted in the autumn sea-

on of 2006. The images were acquired by means of a digitalamera: Olympus C8080 (8 M pixel). This camera had alreadyeen calibrated for previous works. So that it can directly beelated to the realization of accurate metric object restitution.

on the facade of the monuments.

2 ultural Heritage 10 (2009) 269–274

Ortm

tccsuIawicata2

wwttaa

ajsgtiauit

F(

eppom

toaf

•

•

72 G. Karauguz et al. / Journal of C

therwise, it could end up with the problem of detailed objecteconstruction. The choice of this type of cameras is often jus-ified in archaeological contexts by the cheapness, easiness and

anageability in their use.A monoscopic digital photogrammetric station was used as

he photogrammetric system in this project. Such an instrument,apable of determining the intersection of homologue rays, isalled stereoplotter. On the other hand, the photogrammetricystem is based on a software package Photomodeler Pro 4.0,sed for the digital orientation and restitution of photographs.t allows to work with images obtained either with digitalnd/or analogical cameras. The observation system employedas monoscopic. The mathematic model of relation between

mage co-ordinates and ground co-ordinates is the collinearityondition. Orientation parameters and ground point coordinatesre obtained by means of ray beam adjustment. There are twoypes of results: numeric, for the quality control of the results,nd graphic (formats DXF, 3D Studio, RAW, VRML 1.0 and.0, Direct 3D, Wavefront and Iges).

Field work: the control point coordinates (X, Y, Z) (Fig. 1),hich allow to do the orientation, leveling and scaling processes,ere determined with a reflectorless Topcon GPT, 3007 total sta-

ions [25]. The control points were evenly distributed in all overhe site. Subsequently, the topographic equipment describedbove was used to measure the coordinates by means of a closednd orientated traverse made up by six stations.

Taking of photographs: the photographs must be taken inccordance with the principles of photography and they are sub-ect to the requirements of digital monoscopic photogrammetricystems [26]. A large number of photographs were taken (botheneral shots and close-up shots focusing on details), becausehe redundancy of measurements make the detection and elim-nation of gross errors easier. The Olympus calibrated camera

s mentioned above was used. The digital camera was mainlysed to generate the building’s 3D model, as well as for captur-ng images intended to be used as a source for photo-realisticextures.

Fig. 3. The finished work as orthophoto and line artwork (a) a

ig. 2. The finished work as orthophoto and line artwork (a) and 3D modelsfrom side and top) (b) of the monument.

The photos should be taken considering the following points:ach element depicted must be contained in a minimum of threehotos; the convergence between photos taken from differentositions will have to have optimum values of 90o (good valuesf 60o) so that the beam adjustments are carried out well; thereust be at least 50% overlap between photos.Softcopy work: photogrammetric data processing is under-

aken until the scaled-down models of the original structure arebtained. That is why we resort to stereoplotters (software pack-ge Photomodeler Pro 4.0). The processing is divided into theollowing stages:

interior orientation: this operation entails the reconstructionof perspective beams in conditions similar to their formationwithin the photographic camera [27,28]. This operation isdone automatically by the digital photogrammetric station,which introduces subsequent data. In digital camera, lackingfiducial marks, this step involves the introduction of the dataobtained in the calibration of the camera;exterior orientation: in this stage, the beams generated in theinterior orientation process are positioned in relation to the

ground. The position adopted at the moment of exposureof the photographs must remain unchanged. With that pur-pose in mind, at least three known co-ordinate points mustbe marked in each photograph, or six points coinciding with

nd 3D models (from side and top) (b) of the monument.

ultural Heritage 10 (2009) 269–274 273

•

4

iawm

4

rpuitemaeperm

5

dsdmrb

ww

tthamped

btlbaltawOqrTos

sglTiwwstnaoppssptboo

G. Karauguz et al. / Journal of C

another photograph already orientated or to be orientated inthe same process. If support points are available, they shallbe marked during this process as well. Exterior orientationthus includes two operations arranged in sequences, whichare performed automatically by the digital photogrammeticstation, known as [24,27];relative orientation: perspective beams shall be set in a per-spective position. The simultaneous intersection of at leastsix pairs of homologue rays distributed through the model isenough for the remaining points to intersect as well.

.1. Absolute orientation

Once the model has been established, it is necessary to adjustt to the ground’s co-ordinate system. Control point coordinatesre introduced with that purpose. Thus, a new adjustment is donehich grants correct levelling, orientation and scaling of the 3Dodel.

.2. The last step is the construction of the model

The information contained in the photographs will be mate-ialized manually marking all of the elements of interest on thehotographs, so that they can be reflected in the following doc-ments: digital plans (3D and 2D), lists of co-ordinates withnformation on the errors observed, etc. Several tools of resti-ution are available: points, lines, polylines (or other graphicntities such as cylinders, circumferences, etc.). The resultingodels containing metric information and the 3D models are

vailable to be exported in conventional formats (dxf, vrml,tc.) into other programs in order to be visualized, edited orrocessed. Another feature of the system is its capability to gen-rate 3D models of surfaces and the subsequent projection ofeal textures, captured from the object’s photographs onto thoseodels of surfaces (Fig. 2 and 3).

. Conclusion

3D model shown below were produced by the monoscopicigital photogrammetric system used (up to a few years ago, onlytereoscopic analytical systems were available). The proposedigital system works on all kinds of personal computers and theost commonly used operating systems. The low investment

equired to tune the equipment up makes digital systems theest option available [29].

Orthophoto of the monument facade was created by the soft-are used. Then 3D rendering of the monument was createdith photo-realistic textures as shown in Fig. 4.These textures have been implemented starting from the pho-

ographic shots of the object. The results obtained clearly showhe advantages the method has. Information is continuous andomogenous for the whole of the model analysed. There isnother interesting advantage: the amount and kind of infor-

ation to be extracted is a decision that can be made in the lab

hase. Therefore, in this phase, tasks are adapted to each andvery need. This implies greater flexibility, since these actionso not condition fieldwork. On the other hand, one can always go

opos

Fig. 4. Distribution of the camera shooting stations.

ack to the photogrammetric digital files of the building in ordero complete any information missing. Collecting data on a regu-ar basis will enable us to trace the evolution of the pathologieseing studied through time. Digital photogrammetric systemsllow the use of conventional digital cameras and consequentlyower costs (about 600 D ) [30], which is a novelty with regardo analytical photogrammetric systems and photogrammetricnalogical cameras. The shots taken using the digital cameraere directly transferred and processed. The calibration of thelympus digital camera was carried out in order to improve theuality of the numeric and graphic results, determining its aber-ations (focal length, principal point and lens radial distortion).he digital photogrammetric system used also allows the usef non-calibrated cameras. In this case, the system carries outelf-calibration for each of the pictures used.

Due to the characteristics of the Photomodeler Pro 4.0 mono-copic digital photogrammetric system, the tasks required forenerating the 3D models were performed by people with veryittle training, thanks to the user-friendliness of the system [31].he main disadvantage when using a monoscopic vision system

s the slow speed observed in the building restitution process,hen compared to stereoscopic vision systems. This drawbackas overcome in the work described in this paper by carefully

electing beforehand the elements of the building to be resti-uted that would allow a thorough structural analysis. The totalumber of points handled for generating the models was 276nd the total number of photograms 43. Lab work was carriedut by one person for a week. The relative orientation of thehotograms is done after marking in different pictures series ofoints common to at least two shots. Two hundred and seventy-ix points in all were used and special care was taken to makeure that they were homogeneously distributed throughout thehotograms. We decided not to include any more points sincehat would have meant increasing the amount of work with noearing or increase in accuracy. Absolute orientation was carriedut by means of the points measured using topographical meth-ds. Fifty-seven points were used, only those easy to identifyn the photograms, because it was very clear that using dubious

oints made the quality of the results poorer. From this phasenwards, the photogrammetric station enables us to know thepatial co-ordinates of any point and therefore any distance or

2 ultur

mutart

R

[

[

[

[

[

[

[

[

[[

[

[[[

[

[

[

[

[

[

[KB, editor. Close range photogrammetry and machine vision. Caithness,(2001), Scotland, UK: Whittles Publishing.

[31] J. Herráez, P. Navarro, J. L. Cabanes, M. Jiménez, Digital stereo plotting

74 G. Karauguz et al. / Journal of C

agnitude. The smallest possible number of photograms wassed in the process (only 43 of the 62 available were used) sincehe higher the number of photograms, the slower the operationsnd the more tedious and long-winded the overall process ofestitution, since it is necessary to identify manually each pointo be restituted in every picture where it can be identified.

eferences

[1] P. Pegon, A.V. Pinto, M. Géradin, Numerical modelling of stone-blockmonumental structures, Comput Struct 79 (2001) 2165–2181.

[2] W. Bohler, G. Heinz. Documentation, surveying, photogrammetry, XVIICIPA Symposium. Recife, Olinda (1999).

[3] J. Sienz, I. Szarvasy, E. Hinton, M.L. Andrade, Computational modellingof 3D objects by using fitting techniques and subsequent mesh generation,Comput Struct 78 (2000) 397–413.

[4] M.Scherer. About the synthesis of different methods in surveying, XVIIIInternational Symposium of CIPA (2002), Potsdam, Germany, 423–9.

[5] J.R.S. Sterret. Peliminary report of an archaeological journey made in asiaminor during the summer of 1884 (1885), Boston.

[6] H. G. Güterbock. Alte und Neue Hethitische Denkmaler, Halil Ethem HatıraKitabı 1 (l947), Ankara, 48–51.

[7] K. Bittel. Beitrag zu Eflatun-Pinar, Bibliotheca Orientalis 10, l/2 (1953),2–5.

[8] E. Laroche. Eflatun Pinar, Anadolu/Anatolia 3 (l958), 43–7.[9] J. Mellaart, The Late Bronze Age Monuments of Eflatun Pinar And Fasillar

Near Beysehir, Anatolian Studies 12 (1962) 111–117.10] R.L. Alexander, The Mountain-God at Eflatun Pinar, Anatolica 2 (1968)

7–88.11] Behm-Blancke, R. D. Manfred. Ritting, Der ‘Aslantas’ von Eflatun Pinar,

Mitteilungen Der Deutschhen Orient-Gesellschaft 102 (l970); 89–99.12] R. Naumann. Einige Beobachtungen in Eflatun–Pinar’ Mélanges Mansel

(Mansel’e Armagan), 1–3 (1974), 467–74.13] J. Börker-Klähn. Noch Einmal Iflatun Pınar, Aspects of Art And Iconog-

raphy: Anatolia And Its Neighbors. Studies in Honor of N. Özgüc, Nimet

Özgü’e Armagan, (Edited by M. J. Mellink, E. Porada, T. Özgüc) (l976),Ankara, 339–55.

14] Börker-Klähn. J. C. Börker, Eflatun Pinar Zu Rekonstruktion, Deutung undDatierung, Jahrbuch des (Kaiserlich) Deutschen Archäologischen Instituts(mit dem Beiblatt) (und) Archäologischer Anzeiger, 90 (l993), l–41.

al Heritage 10 (2009) 269–274

15] J. Erkanal. Eflatun Pinar Aniti, Bedrettin Cömert’e Armagan, Ankara(1980), 287–301.

16] G. Karauguz. Arkeolojik ve Filolojik Belgeler Isıgında Orta Anadolu’nunGüney Kesimi (2005), Konya.

17] R. Lebrun. À propos de quelques rois Hittites bâtisseurs’ archéologie etreligions de l’anatolie ancienne, mélanges en l’honneur du professeurPaul Naster (édites par R. Donceel, R. Lebrun) (1984), Louvain-la Neuve,157–66.

18] A. M. Darga. Hitit Sanatı, (1992), Istanbul.19] J. Garstang, The Hulaia River Land and Dadassas. A curcial problem in

Hittite geography, Journal of Cuneiform Studies 3 (1944) 14–37.20] I. Singer, Great Kings of Tarhuntassa, Studi Micenei ed Egeo-Anatolici 38

(1996) 63–71.21] J. Friedrich. Hethitischen Gesetze, (l959), Leiden.22] F. Imparati. Le Leggi Ittite, (II 102-2), (1964), Roma.23] A. Almagro. Photogrammetry for everybody, Proceedings of the Sympo-

sium of CIPA in Recife/Olinda (1999), Brasil, 3–6.24] K. Kraus, Photogrammetry. Fundamentals and standard processes. vol. 1.

(2000), Germany.25] http://www.topcon-positioning.eu/index.asp?pageid=9c1dedf719e74e

548650803401d2e683.26] P. Arias. Simple methods for close range photogrammetry. Application to

the study of traditional agroindustrial constructions. PhD thesis. Universityof Vigo, Pontevedra (2003), Spain.

27] W. Zhizhuo, Principles of photogrammetry. Beijing, Publishing House ofSurveying and Mapping, China, 1990.

28] Fryer J.G., Camera calibration’ In: Atkinson KB, editor. Close rangephotogrammetry and machine vision. Caithness, Scotland, (2001), UK:Whittles Publishing: 56–79.

29] P. Drap, P. Grussenmeyer, G. Gaillard. Simple photogrammetric methodswith arpenteur 3-D plotting and orthoimage generation: the image process,XVIII International Symposium of CIPA, Potsdam, (2002), Germany, 47-54.

30] A. Gruen. Development of digital methodology and systems, In: Atkinson

without 3D vision: an effective way in the architectural photogrammetry,XVII CIPA Symposium. (1999), Recife, Olinda, Brazil.