Causes and effectsMethodologies used in digitalization of architectural-urban heritage

Mateusz Pankiewicz11Lodz University of Technology1mateusz.pankiewicz@p.lodz.pl

Since some time already, digital reconstructions in architecture, urbanism andarchaeology are gradually switching from describing built heritage as acollection of static and unchangeable entities towards more compound andexplicit presentation and knowledge management techniques. This includes forinstance data management and multimedia systems, immersive environments orsemantic information modelling such as GIS (Geospatial Information Systems),BIM (Building Information Modeling) or HBIM (Historic Building InformationModeling). Graphical user interfaces, interaction and usability have become anessential part of produced reconstructions. This shift in terms of dissemination ofan architectural and urban heritage that is supposed to increase the socialawareness and participation should be structured in a way that enables recipientsoriginating from different backgrounds to grasp information pertaining to almostany knowledge domain, allowing for self-exploration and interpretation ofpresented knowledge. This paper discusses important nodes of the reconstructionprocess in the spirit of informative modelling that are characteristic for anypossible approach towards conscious heritage representations.

Keywords: Informative modelling, Spatio-temporal modelling, Cultural heritage

CONTEXTIn the field of architectural heritage the objects thatresearchers are striving to reconstruct are never fullyknown. The incontestable fact is that they do havea rich and usually complex history, abundant withevolutions and changes. These changes are usuallyunderstood only in terms of their topology and mor-phology as the physical appearance stresses the pro-cess. Yet the alteration does not affect only the arte-facts materiality. Immaterial aspects of the site areaffected by time passage and events not less. Elusive

cause and effect factors that were influencing the en-tities determining their shape, condition, location orcharacteristics should be taken into consideration aswell. There is a strong need to preserve space for un-certainty as well, as hardly any heritage unit is free ofdoubts emerging from, for instance, heterogeneous,dubious, incomplete or even contradictory historicaldocumentation - if any in the first place. Further-more urban, as well as architectural entities are neverconstant in time as they tend to undergo constantand unavoidable transformations of different types

DIGITAL HERITAGE - Volume 2 - eCAADe 35 | 25

Figure 1Example of densepoint cloudsimplification. Left:Point cloud; Middle:Mesh; Right:Discrepanciesmarked with colourscale. (Source:Garagnani 2013)throughout their lifespan. These events in combi-

nation with plurality of factors that may seem diver-gent or even unrelated at the first glimpse, shapeand form the buildings or whole urban complexesthrough their lifecycles. Moreover, the typologicaldifference between architectural and urban entities,which results from the fact that the urban environ-ment can be perceived as a container for multiple ar-chitectural components, awakens the awareness ofdifferent granularities - or scales - of aforementionedcausative factors and events. Better or worse pre-served remnants of architectural past encounterednowadays are therefore resultants of all these pro-cesses that had occurred. Amalgamate of these fac-tors can be eventually perceived as the construct ofwhat could be called the meaning of the built arte-fact.

INFORMATIVEMODELLINGComplex, multi-layered network of intersecting ele-ments and types characterizing architectural artefactrequires adaptive representation model. In the do-main of architectural heritage gathering, analysing,structuring and understanding various types ofsource documentation is the fundamental part of theprocess, whereas visualizing and retrieving informa-tion is the paramount aim of each. Therefore infor-mation visualization - where the 3D representationof an architectural entity serves only as a perceptivetool and does not constitute the final result of the re-construction process - seems tomore accurately fit inthe general idea. This was summed up precisely by E.R. Tufte and J. Bertin who said respectively that: ”Weenvision information in order to reason about, com-municate, document and preserve that knowledge”

(E.R. Tufte 1997) and ”...a graphic is never an end initself: it is a moment in, a process of decision mak-ing...”(J. Bertin 1967). This is already the basis of theinterdisciplinary methodological approach - or con-cept - of an informative modelling in which the rep-resentation of artefacts does not claim veracity, butsupports dynamic information retrieval and visuali-sation, reasoning and cognition. Abstraction (the in-formation visualisation legacy) and figuration (the ar-chitectural representation legacy) are integrated asalternative/mixable types of representation, allow-ing partial knowledge to be communicated and im-portant notions in historic sciences such as data un-certainty to be conveyed graphically. Though arte-facts’ accurate and precise representation is not de-termined, present advancing data-processing tech-nologies allow for more comprehensive and massivedata acquisition from heritage sites. These becomenecessary in order to describe, understand and sup-port built heritage. Moreover developing softwaretechnology allows to combine models acquired dur-ing high definition surveys with the informative lay-ers, as well as to adopt them for logical re-use in var-ious, process depending scenarios. Nevertheless theveracity of the model remains still in question as thedense point clouds need to be discretised into moreefficient and reuse-ready meshes or polygons. Thisin turn demands simplification of the acquired dataset, which results in some level of divergence fromthe original. (Figure 1) Authenticity of the model ishowever, out of the question when we discuss mod-els constructed solely on the basis of existing his-toric documentation, which fundamentally must betreatedwith cautionandcannotbe taken forgranted.

26 | eCAADe 35 - DIGITAL HERITAGE - Volume 2

BI-DIRECTIONALITYWith progressive advancement of information tech-nologymore andmore options to store, manage andpresent acquired datasets are available for the re-searchers. Nevertheless it occurs to the author thatit is currently possible to determine two particular di-rections in which the process of digitizing built her-itage is heading. The foundation for this specific bi-directionality lays on the one hand in the develop-ment of BIM theory into practice which results - froma CAD (Computer Aided Design) point of view - inthe release of dedicated specific software such as Au-todesk Revit or Graphisoft Archicad, as well as prolif-eration of its assumptions through diverse groups ofinterests such as architects, archaeologists, historians- to name a few; and on the other hand in the needto search for and create custom solutions dedicatedto particular problems or issues, which is strongly en-couraged and motivated with the spread and popu-larisation of programming languages and highly ad-vanced frameworks such as Unity 5 or Unreal 4 en-gineswhich from some time already are no longer re-served only for a top-end professionals or AAAdevel-opment studios. While both approaches are concep-tually different they try to serve the same purpose:to gather, manage and disseminate the knowledgeabout built heritage.

OUTSETRegardless of which approach is to be chosen finally,there are still some strategic decisions to be madeand steps to be fulfilled initially. This article aimsto describe them in more detail. It highlights threefactors that are fundamental for the reliable organi-zation of such a complex multidimensional recon-struction model: addressing temporal issues, spatio-temporal representation of uncertainty in data, hy-potheses and a spatial organization of the model toallow diverse usability. Source data handling is notgoing to be included in this paper as this particu-lar matter rather depends on their quality, quantity,type and specific purpose of undertaken reconstruc-tion process. Therefore its management and treat-

ment may vary and be performed in toomany differ-ent ways using multiple approaches exclusively or inparallel.

Figure 2Possible types ofchange of aspatio-temporalobject

THE OCCURENCE OF CHANGEAs it was already mentioned, every built entitychanges during its lifespan. Number of possibletransformations it could go through are vast, thoughit is possible to distinguish most relevant: build-ings are built, destroyed, rebuilt, they may be ex-tended, attached to another entity or divided intoseveral parts. Eventually they can peacefully degradethrough the whole life-cycle. Last but not least itcan occur that the building would be totally or par-tially erased and rebuild in the same or different lo-cation (e.g. city of Warsaw after II World War). Thisstatement makes apparent that whenever there is achange in space (spatial alteration) it never happensinstantly but instead implies that some time passedfrom the beginning of transformation to its end. Un-deniably time plays an important role in the descrip-tion of change. Even more when we take into con-sideration that not only spatial form of the buildingis changing but also its relationship with surround-ing entities or its particular attributes. Reassuming,the change can affect objects morphology, topol-ogy or attribute definition and it never lacks a tem-poral dimension. (Figure 2) The time of each alter-

DIGITAL HERITAGE - Volume 2 - eCAADe 35 | 27

ation is, as a matter of fact, never constant. Eachtype of change lasts different amount of time. For ex-ample the demolition of the building can take just afew days, while its construction could last for yearsor even decades. This implies that time of the eventoccurrence is therefore scalable., which means thattemporal incidents have different granularities. Suchcomplex spatio-temporal transformation system oc-curring on different - usually nested - levels with var-ious scales need a proper and suitable data manage-ment system. Pelekis (Pelekis et al. 2004) describedandcomparedeleven ready-to-adoptdatabasemod-els with various levels of complexity, each accuratefor certain tasks and aims. The question of whichmodel to choose relies mostly on the defined goalof the undertaken reconstruction, its intricacy, querystructure and operations it should be capable to per-form simultaneously. Whereas somemodels are sim-ple and operate on the snapshot or time-stampingbased structure (Figure 3), others reflect quite elabo-ratedmechanisms derived fromgraph theory (Figure4) or Object Oriented Programming (Figure 5).

UNCERTAINTY ISWHAT REALLYMATTERSIn the domain of digital heritage a lot of effort is putto create accurate and precise digital reconstructionmodels from available spatial data, preferably usingTerrestrial Laser Scanning or photogrammetry tech-niques supported by historic documentation. Simul-taneously it limits the scope of the reconstruction tothe digitalisation of only what is left or, perchance,to particular states of the building to which availabledocumentation does not arouse any level of uncer-tainty or suspicion. However, in the field of archi-tectural heritage, more than anywhere else, due toits unstable and long-term character, not only spa-tial, but temporal aspects need to be taken into con-sideration. Traces in form of historic documenta-tion stored in archives (if there are any in the firstplace!) are rarely satisfying the researchers. Het-erogeneous, dubious, incomplete or even contradic-torydocumentation leavesmanyquestionmarks anddoor open for hypothesis and introduces the pos-

sibility of uncertainty concerning spatial, as well astemporal or even attributable aspects.

Figure 3Graphicapproximation ofexemplaryspatio-temporaldatabase models:SimpleTime-Stamping(Source: Pelekis etal. 2004)

Figure 4Graphicapproximation ofexemplaryspatio-temporaldatabase models:History GraphModel, (Source:Pelekis et al. 2004)

Figure 5Graphicapproximation ofexemplaryspatio-temporaldatabase models:TheSpatio-TemporalObject Model(STOM) (Source:Pelekis et al. 2004)

Spatial uncertainty results usually from lack of mate-rial on which the researcher could base the recon-struction. Historic sites were rarely, or never, sur-veyed in the past when no one would expect themto be historic one day. Most of the archived mate-rial consists usually of crude designs and drawings,but almost never of post-construction surveys. This

28 | eCAADe 35 - DIGITAL HERITAGE - Volume 2

makes researchers working with almost hypotheticsource documentation, as the changes made on-site during construction of the building were, withthe high level of probability, never marked or de-scribed. Researchers are usually left with pieces of in-formation on which they are supposed to build theirproverbial church. As it is impossible to find just oneideal solution that would describe changing build-ing morphology it is crucial to refer to the hypothe-sis based on reliable sources not necessarily concern-ingparticular entity but at least similar from the sametime period. Temporal uncertainty, on the otherhand, is even more complex and can be describedwith several factors. First, and probably one of themost important, is the problem of dating exactness,which reminds instantly about time scale and granu-larity. Eventswhich tookplace in XIXth century, in year1845, between the year 1897 - 1909 denote differenttemporal weights. If the scale is the century - is therea possibility to define the year precisely? Or if thescale is the year - is it possible to define the month?This imprecisions in dating description causes lack ofcohesion andpresents first important technical prob-lemaswell -which constitutes the timeline andeventpositioning. Another trouble arises with the inter-pretation of the heritage sites. The past is usually in-terpreted by different kinds of historians (the histori-ans, art or architecture historians, etc.) on the basis ofavailable historic sources. This can undoubtedly leadto a formulation of various and divergent interpreta-tions - or rather - hypotheses concerning particularentity, site, event or period of occurrence. As a resulteach formulated opinion is laden with some level ofuncertainty, even despite the purest intentions of theinterpreters, which in turn affects the entity and pos-sibly distorts its spatio-temporal change pattern de-viating it from unfortunately unknown reality. Real-ising the potential of the stratified uncertainty couldleave us with the thought that the only real and cer-tain element of spatio-temporal reconstruction chainis the existing residue of an object and that every at-tempt to model its past state is resulting in creatingjust a hypothesis.

STRUCTURAL SYSTEMATICSTo entirely utilize the potential of building entitiesfor the sake of performed reconstruction it is nec-essary to classify them in order to withdraw max-imum spatial information from general data. It ispossible to distinguish three phases of such ordina-tion. First one assumes classification according to theadopted point of view - which means distinguishingelements according to their e.g. function or mate-rial they are composed of, as well as to other crite-ria such as a time of erection or style they were builtin, etc. Second differentiates elements according totheir morphological decomposition in the life-cycleprocess. Eventually, in the third phase associationsamong concepts are created and visualized.Workingwith historic structures requires constant reasoningabout temporal changes. This in turn necessitatesproper model structuring and reorganization whichleads to the question of its morphology as well as itsattributes and spatial relationships with surroundingentities. Recalled earlier, the concept of granularity,has a strong influence not only in case of temporaldimension but spatial and structural as well. Eachbuilding undergoes changes, but each one is a com-position of parts which, as a matter of fact, also un-dergo change. Accordingly, thisworks in opposite di-rection: a building is just an element of a larger groupof buildings which in turn constructs a neighbour-hood, district or town. Therefore it is possible to in-troduce various levels of spatial granularity as well.Defining this hierarchy is important, as every ele-ment or entity would have its own specific attributes,sometimes utterly unique, sometimes shared withothers. In general it is possible to organize objects inthree nested levels: groups - agglomerating buildingcomplexes, single buildings and their major compo-nents; entities - corresponding to the functional andtemporal model divisions; references - representingsome specific aspects of extracted entities. This pe-culiar morphological inception could be than limitedby setting theminimumandmaximum level of detailthat seems sufficient for thepurposeofperformed re-construction.

DIGITAL HERITAGE - Volume 2 - eCAADe 35 | 29

CONCLUSIONThe society of information we are currently experi-encing, demands an informative, scientific approachto each and every aspect of life. This is particu-larly valid and important in the domain of architec-tural heritage, which is built on the knowledge ofthe past. Therefore the number of methodologiesor approaches to the matter of architectural-urbanreconstruction processes seems fairly limited, as nomatter what, any of them would revolve around theidea of creating a knowledge management systemconcerning the heritage entity. Research performedfor this paper revealed that all analysed case stud-ies did encounter particular problems while concep-tualizing systematic workflow. Spatio-temporal datamanagement, insufficient sources that provoke andfuel uncertainties related to spatial, temporal, as wellas attribute layers, handling of existing hypothesesare mostly reported. Therefore the gravity centre ofthis article was moved a bit from depicting method-ologies towards describing these specific challeng-ing nodes that in the same time determine the struc-ture of informative reconstruction model. Descrip-tion and potential guidelines for handling such caseswere formed on the basis of available literature. Ap-plicationof thesenodes in theprocess of reconstruct-ing architectural or urban entities of different scalesseem valid and critical for implementation either inBIM driven reconstruction models or custom-madesolutions.

REFERENCESAnderson, EF, McLoughlin, L, Liarokapis, F, Peters, Ch,

Petridis, P and de Freitas, S 2009, ’Developing seri-ous games for cultural heritage: state-of-the-art re-view’, Virtual Reality, -(10), pp. 255-275

Bertin, J (1967) 2011, Semiology of Graphics, Esri Press,Redlands, California

Dudek, I and Blaise, JY 2005, ’From artefact represen-tation to information visualisation: Genesis of in-formative modelling’, in Butz, A, Fischer, B, Krüger,A and Olivier, P (eds) 2005, Smart Graphics, LectureNotes in Computer Science vol. 3638, Springer, Berlin,Heidelberg, pp. 230-236

Fai, S, Graham, K, Duckworth, T, Wood, N and Attar, R2011 ’Building Information Modelling and HeritageDocumentation’, Proceedings of XXIIIrd InternationalCIPA Symposium, Prague, Czech Republic

Garagnani, S 2013 ’Building Information Modelling andreal world knowledge. A methodological approachto accurate semantic documentation of the builtenvironment’, Digital Heritage. Proceedings of Digi-tal Heritage International Congress, Marseille, France,pp. 489-496

Kimberly, D, Frischer, B and et, al. 2010 ’Rome Reborn2.0: A Case Study of Virtual City Reconstruction Us-ing Procedural Modeling Techniques’, Making His-tory Interactive. 37th Proceedings of the CAA Confer-ence, Williamsburg, Virginia, USA, pp. 62-66

De Luca, L, Busayarat, Ch, Stefani, Ch, Véron, P and Flo-renzano, M 2011, ’A semantic-based platform for thedigital analysis of architectural heritage’, Computers& Graphics, 35(2), pp. 227-241

Di Mascio, D, Chiuni, M, Fillwalk, J and Pauwels, P 2016’3D Digital Reconstruction of Lost Buildings’, Pro-ceedings of the 34th eCAADe Conference, Oulu, Fin-land, pp. 511-520

Murphy, M, Mcgovern, E and Pavia, S 2013, ’Historicbuilding infromation modelling (HBIM)’, ISPRS Jour-nal of Photogrammetry and Remote Sensing, -(76),pp. 89-102

Pelekis, N, Theodoulidis, B, Kopanakis, I and Theodor-idis, Y 2004, ’Literature Review of Spatio-TemporalDatabase Models’, The Knowledge Engineering Re-view, 19(3), pp. 235-274

Renolen, A 1999, Concepts and Methods for ModellingTemporalandSpatiotemporal Information, Ph.D. The-sis, NTNU

Stefani, Ch, De Luca, L, Véron, P and Florenzano, M 2010,’Time indeterminacy and spatio-temporal buildingtransformations: an approach for architectural her-itage understanding’, International Journal on Inter-active Design and Manufacturing (IJIDeM), 4(1), pp.61-74

Tamborrino, R andRinaudo, F 2015 ’Digital UrbanHistoryAs An Interpretation Key Of Cities’ Cultural Heritage’,Digital Heritage. Proceedings of Digital Heritage Inter-national Congress, Marseille, France, pp. 235-242

Tufte, ER 1997, Visual Explanations, Graphic Press,Cheshire

30 | eCAADe 35 - DIGITAL HERITAGE - Volume 2