According to the analyzed literature, 3D virtual modeling in the architectural heritage field is done in several ways and by using different approaches. These differences allow for research to be grouped according to its purpose, degree of automation, degree of segmentation, etc. Therefore, the bibliographic review that is described below has been subdivided into categories according to the tools used in each. The first category corresponds to those researchers who have adopted only the commercial BIM platforms to develop BIM “as-is model” models, and in turn create the parametric H-BIM libraries.
The analysis of the current literature on the virtual modeling of existing buildings has demonstrated the benefits of the use of BIM and H-BIM for the preservation, management, or parametric modeling of cultural heritage (CH).
5.1.1. First Category: only BIM Methodologies
In this category, thirteen documents that present case studies to achieve virtual and parametric 3D modeling of the buildings that are analyzed have been studied. From these thirteen documents, ten have adopted the Autodesk Revit software and three the Graphisoft Archicad software to illustrate the implemented methods. In turn, four documents create H-BIM parametric libraries and the other ten only develop the complete parametric model of the analyzed heritage. Despite the BIM tools used and the purpose of their results, the methodologies that are proposed in the most representative documents of this category have particular similarities for performing such actions as cuts, sections, and levels on the acquired point clouds, which allow for the parametric modeling to be accelerated and optimized. Most of the analyzed works do not declare the degree of automation or the diagnostic and management processes implemented.
For Murphy et al. [15
], the process of creating an H-BIM library begins with the 3D modeling of parametric objects through the integration of point clouds, photogrammetry, historical documentation, and Graphisoft Archicad software. File format and data exchange problems are fixed with the script GDL language included in this software. The simple or regular solid objects of the analyzed heritage are modeled using the simple primitives that are provided by the platform. On the other hand, the irregular shapes found from the analyzed heritage monuments are modeled in two parts. Primarily, a series of sections made on the point cloud are created to identify the profile of each object and the surfaces are subsequently converted to solid elements through NURBS, mesh, and Boolean operations (included in the GDL).
The methodologies that are implemented in these approaches increases the speed of the parametric modeling of the analyzed heritage. However, they do not accurately develop the steps to segment the point clouds, generating confusion in the different conservators who try to implement this architectural modeling process.
The above is also acknowledged in the article by López et al., 2017 [100
], in which a three-step methodology is developed to create an H-BIM library of the Romanesque church of Santa María la Real de Mave, Palencia, Spain. The first step covers the collection of semantic and spatial information. In the second step, the obtained data are processed and organized for feeding into the BIM Autodesk Revit platform. In the third step, a design of multiple views, cuts, sections, and grids made on the point cloud are created, following the rules and the constructive patterns of the architectural period that the building belongs to.
As in the previous case, the uniform and simple surfaces are modeled using the basic tools of BIM environments. On the other hand, irregular or complex surfaces are modeled in two stages. The first stage starts with the sketch of the current component profile on a 2D work plane. Subsequently, the reference planes, 2D profiles and modeling tools are integrated to create the solid elements in a 3D work environment. The H-BIM Library developed allows the maintenance, management and rehabilitation of historic buildings. In this approach, the semi-automatic implementation is time consuming and it represents a certain weakness.
Such authors as Del Giudice & Osello, 2013 [101
] and Biagini et al., 2016 [102
], have also developed a process of virtual modeling in their methodologies based on the design of multiple sections, cuts, levels and grids made on the point cloud. However, they differ from the above-mentioned articles because they present a new approach to the temporal calculation in the Autodesk Revit platform. This approach is effective for the order and control of the architectural heritage restoration and management projects. These documents have not accurately developed the necessary steps to model and manage each parametric object, and neither have H-BIM libraries been developed, which represents a problem and a disadvantage for their implementation.
There are other important works that could be mentioned in this category. Brumana et al., 2013 [103
], Fai & Sydor, 2013 [104
], and D’Auria et al., 2014 [105
] have all used a similar workflow to link point clouds with the BIM Autodesk Revit platform. Their main topic is to define or to separate the components to be modeled, depending on their typology, hierarchy, and material. This separation facilitates the exchange of data between the different virtual modeling software. The authors argue that the lack of flexible tools in BIM generates drawbacks for historic buildings modeling.
Such authors as Ma et al., 2015 [106
] Achille et al., 2015 [107
], Cheng et al., 2015 [108
], and Adami et al., 2016 [109
], describe how the modeling of historical components with BIM platforms can help researchers to visualize a representative 3D model, so as to know the real information of each object, as well as its implementation in the maintenance, management, and restoration processes of the monument. These works have not developed a methodological process on virtual modeling, which impedes its implementation by other researchers to create effective workflows.
These publications about BIM modeling for existing buildings, or about H-BIM, have shown that software tools have not yet been developed to automatically convert point clouds into BIM components. In addition, the lack of algorithms to automatically convert the complex shapes of the point clouds into BIM parametric objects is another bottleneck in the field of CH reconstruction.
Another common limitation or difference in the comparison between the methodologies analyzed is that each one is often designed to highlight the innovation of a particular method over the others, in some cases, provoking gaps in the information that they want to transmit.
5.1.2. Second Category: BIM, Auxiliary Tools and OSS Methodologies
Currently, BIM platforms are not able to solve all of the issues that come from the modeling management and the maintenance of the architectural heritage; therefore, the integration of other instruments or auxiliary tools is needed. In this category, a total of twelve documents have been analyzed. From the twelve documents, seven create H-BIM libraries and five do not. In turn, five papers adopted the commercial Autodesk Revit platform in combination with Rhinoceros and its plug-in Non-Uniform Rational Basis-Splines (NURBS); while, the other four adopted the same platform in combination with BIM Open Source Software (OSS), Autodesk AutoCAD, and 3DMax or Blender. On the other hand, two authors combined the commercial BIM platform, Graphisoft Archicad, with OSS, and the application-programming interface (API), while one author combined Autodesk Revit with its API programming interface. The use of external tools for the efficient and accurate modeling of the analyzed building is the differential factor between these categories.
A workflow of the virtual modeling of the Four Courts in Dublin-Ireland, using the architectural historical documents, physical-constructive analysis, and point clouds, is proposed in Dore et al., 2015 [110
]. Parametric objects that are created using the Graphisoft Archicad software and the GDL (3D “RULED”) function are part of an H-BIM library. This approach proposes two methodologies: The first consists of modeling the elements using the historical documents and the second is designed for the accurate and the efficient modeling of irregular elements.
For the second methodology, a design of multiple sections, cuts, and levels that were made on the point cloud is created with the Graphisoft Archicad software. The sections that were made on the point cloud are diagrammed as closed, sized polygons using lines and arcs, and serve as a guide to the 3D “RULED” function to generate accurate BIM models from the point cloud.
On the other hand, in Nieto et al., 2016 [111
], an innovative processing, cataloging, analysis, and identification methodology of the information is proposed, which begins with the creation of a set of grids on the surface under analysis. For this reason, a set of instructions and algorithms that were written in the API of the Graphisoft Archicad software was created in Nieto et al., 2016 [112
], so as to accelerate the automatic generation of the standardized vector grids. The resulting grids serve as a basis for the management, delineation, and recording of the information that can be used to develop an efficient, accurate, and parametric representation of the analyzed architectural heritage. This article is suitable for research focused on identifying and cataloging processes of elements.
Other authors, such as Oreni et al., [112
] and Barazzetti et al., [115
], argue that BIM software presents limitations for modeling non-standard architectural objects, characteristic of architectural heritage. Therefore, the integration of the Rhinoceros software is proposed to model the architectural components of the analyzed building and to generate the H-BIM library. These authors develop a solution that is based on performing profiles on the point cloud from the use of cuts and levels. These cuts and levels serve to interpret and to separate the different profiles extracted from the point cloud into regular or irregular surfaces.
Regular surfaces are modeled with the Revit software tools and irregular ones are modeled thanks to the vector profiles that were made by NURBS curves, a number of algorithms used in the Boolean operations and the Rhinoceros software. The results that were obtained are exported to the Autodesk Revit platform to obtain parametric models that will be part of an H-BIM library.
Another approach for accurately building modeling is presented in Quattrini et al., 2015 [117
]. In this approach, a methodology that minimizes the number of steps and accelerates the modeling of the parametric elements through the integration of point clouds with the Autodesk Revit software is developed. In this document, point clouds are considered to be an excellent source of information, and therefore, cuts, sections, and levels are not performed on the point cloud: each component is modeled directly on the point cloud, thus maintaining the quality and accuracy of the obtained data and details.
This approach classifies the point cloud into regular or irregular surfaces. Regular surfaces are modeled with the Autodesk Revit parametric element libraries and complex ones are created in Boundary Representation (B-Rep). Finally, the Open-Source plug-in Protégé is used to facilitate the integration of parametric data with each modeled element.
Other works could also be mentioned in this category, such as Fregonese et al., 2015 [118
] or Rodríguez-Moreno et al., 2016 [119
]. In these approaches, the integration of point clouds and historical data with a series of specific software (AutoCAD, SketchUP, 3DStudio Max) and a set of Open-Source Software (OSS) (3DReshaper, BIM3DGS) are used to obtain historical virtual models. In addition, the models that were obtained are exported to the Autodesk Revit software to obtain parametric information, thus enabling the resulting model to help manage and conserve existing buildings. In addition, Fregonese et al., 2015 [118
] uses the BIM3DSG software for the management and administration of the model.
Another approach that describes the use of parameterization as a process for contributing to the general framework of “Smart heritage” is presented in Rua et al., 2014 [120
]. A methodology is proposed consisting of five main phases: data collection, architectural study, data processing, digital modeling, and virtual modeling. Autodesk AutoCAD and ArGIS software are used to process technical and geographical data. On the other hand, Autodesk Revit tools are used to model, manipulate, and adapt families that are located in the BIM library to the project and Blender is the software used for the visualization and interaction between the virtual model and the general public. The resulting model is a database capable of being used for the management of the analyzed CH.
Another author that emerges from the cultural heritage community is Garagnani, 2013 [121
]. This author has developed a plug-in called GreenSpider in the Autodesk Revit software API. This plug-in is able to recognize and select the most determinant points of a point cloud and translate them to native reference points in the Revit modeling environment. In addition, GreenSpider has the ability to connect the points and interpolate the vertices through splines, in this way, allowing for explicit parametric representations of the actual captured surfaces to be generated. After its modeling, the components are fed, through IFC, with the semantic and topological metadata that was stored in a BIM file.
The disadvantage of the approaches that were analyzed in this category is that the lack of intelligent algorithms capable of automatically converting point clouds into parametric objects makes the constructive process of irregular components a challenging and time-consuming process. In addition, the architectural elements are modeled on point clouds using intermediate software, which could result in information being lost when exporting the said elements, and therefore in the conception of the building’s general perspective.
5.1.3. Third Category: BIM, Auxiliary Tools, OSS and GIS Methodologies
BIM platforms with geometric/semantic and GIS software are combined by five authors to create detailed historical models. All of the documents create H-BIM libraries. In addition, three documents that were adopted Autodesk Revit and Rhinoceros in combination with the GIS Autodesk InfraWorks software, while another paper adopted Autodesk Revit and Graphisoft Archicad in combination with the GIS (SIGEC and SICaR) software, and another combined Graphisoft Archicad with Sketch-up and their CityGML complement.
Baik et al., [122
] and Baik, 2017 [124
] propose integrating point clouds, Islamic historical manuscripts, and the Hijazi architectural patterns with Autodesk Revit platform, Rhinoceros software, and the Autodesk InfraWorks GIS system, in order to achieve a detailed virtual reconstruction of the Jeddah Historic BIM (JHBIM) library. This library has been specifically designed to accelerate the modeling of Jeddah historical monuments.
First, the point cloud is divided into main-parts and sub-parts. Then, the Autodesk Revit software is used to model the simple objects of the project. On the other hand, irregular surfaces are modeled using NURBS tools of Rhinoceros software. Finally, the BIM 3D model and the 3D GIS Autodesk InfraWorks system are integrated to obtain an accurate and detailed virtual model. The methodologies analyzed have some disadvantages: in particular, because the virtual modeling is a time-consuming manual process and the intervention is limited to Islamic architecture.
Oreni, 2013 [125
] proposes integrating metric and semantic information that was obtained from the GIS systems (SIGEC and SICaR) and point clouds, with GraphiSoft ArchiCAD and Autodesk Revit platforms, in order to create an H-BIM library that would provide the basis for the restoration, conservation and management of heritage buildings, and also to facilitate the interoperability with other interested conservators. Although GIS are today the most agile tools to collect, manipulate, and manage the different data on buildings, it should be noted that this article does not effectively develop the steps that are required to integrate GIS systems with BIM platforms, making it difficult for researchers to implement this methodology as a guide to obtain 3D virtual models.
Dore & Murphy, 2012 [93
] propose using the GraphiSoft ArchiCAD GDL scripting language to achieve the 3D virtual modeling of the parametric objects. Later, the objects that already exist in the internal library are combined with the modeled elements and a set structure is created that is incorporated into the design of an H-BIM library. Finally, the 3D model can be integrated into a GIS environment (ArcGIS) using Sketch-up, with its CityGML add-in. The resulting model serves as the basis for the analysis and the management of the information required for the maintenance of heritage buildings, as well as to accelerate the modeling process of other historical buildings with a similar architectural style.
Although, in this category, the authors have demonstrated standardization in the performance analysis and development of the parametric elements, the ’receptor scenarios’ have shown that for our contribution, the gaps and variety of their implementation can generate disadvantages for modeling a particular building or monument.
5.1.4. Fourth Category: Methodologies without BIM
Finally, BIM platforms are not used as a basis for historical heritage modeling, management, and maintenance by some authors, because they consider that the common BIM software is still unable to manage the huge quantities of data coming from laser scans or photogrammetric surveys, which makes the use of external free or open intermediate software tools (F/OSS) inevitable. In these documents, H-BIM libraries have not been developed. In this category, six documents have been analyzed; two adopted Autodesk AutoCAD software and some OSS software as auxiliary tools, while three other authors create their own tool to model the analyzed heritage. In addition, one document has adopted Rhinoceros in combination with the OSS software (BIM3DSG) and the API of the program.
Fassi et al., 2015 [126
] have developed a methodology for the historical 3D modeling, using the Rhinoceros NURBS in combination with an OSS (BIM3DSG) and a series of algorithms written in the program API. In this approach, the main topic is to develop an (OSS) easy-to-use and easy-to-learn system, where 3D models can be easily visualized and manipulated using any common touch-screen device. The lack of a descriptive methodology about the use and implementation of the new system, as well as the necessity for continuous Internet connection, are the weakest points of this approach.
On the other hand, Soler et al., 2017 [127
], have developed software that is based on GIS systems called Agata. This software is capable of storing and managing all kinds of information associated with each modeled architectural heritage element. It uses (eXtensible Markup Language) XML as a data exchange format. In this approach, the main topic is of great interest for the implementation, but the proposed methodology does not accurately develop the operation and utilization of the new software, or their interoperability with other programs.
In their article, Aguilera & Lahoz, 2010 [128
] have developed a flexible and low-cost system for virtual 3D modeling of archaeological sites. In this document, triangulation algorithms that are based on the 2.5 D Delaunay Triangulation approach are used to convert the image into a polygonal mesh model and to automatically model the elements using polygonal boxes. Subsequently, the models are textured using the Z-buffer algorithm. The resulting models can be grouped and directed to the creation of prototypes to encapsulate and reuse any set of objects. The main drawback of this methodology is that it presents difficulties for the 3D modeling of irregular objects.
The approach of San Jose et al., 2013 [129
] develops a new software platform for the volumetric visualization of complex architectural objects and their application to the teaching, management, and conservation of architectural heritage. The proposed methodology intends to incorporate the knowledge of experts through an open-source platform that is based on GIS systems. Such authors, such as as Guidi & Russo, 2011 [130
] or Micoli et al., 2013 [131
], propose the integration of point clouds and historical documentation with AutoCAD 3D software and OSS tools to increase the real knowledge of a heritage building and support the historical interpretation. First, point clouds are used as a reference for CAD modeling and as a source of direct information. Later, a web application is used to allow for real-time navigation and achieve the access and management of data in a tridimensional context. The resulting model should serve as a 3D data repository.
As in the previous cases, the lack of a detailed methodology to describe the modeling process of each component is a problem and a disadvantage when being used as a guide by other researchers. It is also worth noting that the use of OSS and API has grown significantly in recent years. However, there are no OSS platforms that are able to integrate all of the stages of the documentation and the modeling process of the heritage components. Another disadvantage of this category is that they fail to develop parametric models. Therefore, BIM software would be needed for the parametric modeling of the different architectural components.