The Via Consolare Project in Pompeii  
  Home     Research     Internships     Team     Links     Contact  
  Summary     2005/2006     2007     2008     2009     2010     2011     2012     2013     2014     VCP 3D 
Collaboration and Reconstruction
Amongst the most important priorities of the Via Consolare Project is that our work should serve to facilitate and promote scholarly collaboration between the numerous projects of archaeological research within the ancient city, and to produce results that may be extended by other researchers, easily migrated to newer formats, and put to continuing use for future research. To this end, the Project focuses on the the provision of reusable, high-quality, 3D data resources using open-source and freeware software methods that produce largely open format or easily transferable end products. Hopefully, this will encourage the formulation of city-wide connections and collaborative interpretations, helping to promote data sharing and longevity, as well as allowing for the pooling of academic resources toward the ultimate goal of a unified and comprehensive information system for the site of Pompeii. In addition, the Project has worked to test and standardise open-source, free, and lowest-cost solutions so that these methods may be implemented by projects of all sizes and budgets, not only the wealthy few.

Structure from Motion: Inexpensive 3D Scanning
The latest open-source 3D data acquisition solution that has been incorporated by the Project involves the use of software implementing a structure-from-motion methodology to produce dense surface models of archaeological features and capture colour and 3D positions, often with millions of points per model, that preserve the precise position, arrangement, topography, and appearance of each stone, potsherd, brick or feature. Once geo-positioned, these point clouds and triangulated meshes provide an unparalleled record of the standing remains and even of excavated deposits in situ, permitting precise production of illustrations, phase plans, and complete 3D reconstructions. The advantages of producing such records in archaeology has long been made obvious by the use of time-of-flight or structured-light laser scanners to record standing remains and even excavations in process. What makes this method especially exciting is that it employs only low-end consumer-quality digital cameras and average quality computers in its data pipeline. Primary processing can be done on field laptops in order to ensure that appropriate images have been taken in the field before full processing in the off-season. Furthermore the method is entirely scalable, and may be used to capture the details of small finds or small features, as well as whole insula frontages, in the desired level of resoltuion.

Point cloud of a wall in Insula VII 6

Wireframe suface mesh of the excavated trench
Hugin input photo prior to ortho-rectification

Image after ortho-rectification with comparative survey outline

Hugin and Automatic Stitched Rectified Photography
During wall analysis, the Project spends a great deal of time examining the preserved remains closely and from all angles, and reconstructing a logical sequence of events from the small traces of construction events that each wall contains. A vital step in this process is to document every wall surface, both in detailed notes, normally accompanied by a sketch, photographs and finally a scaled drawing of the wall and the stratigraphic evidence that it preserves. These elements form the primary data collected in the field and enable theories of building sequence, phasing and urbanisation to be developed, and later put to the test by stratigraphic excavation. In Pompeii, simple photography of the wall surface is often insuffient for this process because most wall surfaces cannot be photographed without significant perspective distortion; it is often impossible to move far enough away from the walls surface to take a complete, head-on photograph. This is especially the situation in long, narrow corridors and tiny back rooms, and the problem cannot be solved by simply taking many close-ups, as this does not provide an accurate, consistently-scaled record. Traditionally, these difficulties have been overcome by hand-illustrating architectural sections by means of offset measurements. However, given the roughly 1735 wall surfaces in our research areas, it was clear from the start of the Project that a digital solution might serve to facilitate this process.

Perspective correction, ortho-rectification and automated stitching of digital photographs using Hugin and Panorama Tools has proven to be both considerably faster and more accurate than traditional, pencil and paper methods, producing an end-result that not only records the particular details of each surface in much greater detail (the particular coloration, position and texture) but also achieves this in roughly 1/20th of the time required by traditional methods. Automatic stitching of images means that super-high detail images can be produced, well beyond the capabilities of even professional level digital cameras, including traditional panoramic shots, but also cut-through sections and photographic elevations of standing remains. Furthermore, this method has a number of distinct advantages in comparison to 3D scanning and close-range photogrammetry, being certainly more cost-effective, requiring only a simple digital camera, basic tools for temporarily and non-destructively marking the wall surfaces, and free software, which can run easily on low-end and low-cost computer systems (especially the portable systems we tend to use in the field). There is also little equipment to transport and maintaining the necessary devices in field conditions is not taxing. The end product is an easily archivable digital image which can form a component of the permanent record, or provide the basis for digital inking of outlines and tracing the phases as sketched in walls field notebooks during the original analysis.

Though this methodology is in many cases bound to be supplanted gradually by the 3D point cloud aquisition discussed above, it still has a number of distinct advantages that mean it will continue to be used by the Project as one component of a broad and flexible digital toolkit. Notably, it still forms the most useful way of flattening and recording wall decoration, and is especially useful in conjunction with 3D data capture methods for providing surface decoration textures in even higher resolution than possible from point clouds alone (which can sometimes be hampered by solid, single tone areas). Overall, the goal has been to produce a robust and scalable system for efficiently producing the type of records necessary for systematic and on-going study of complicated architectural features and the development of a working pipeline that will produce publication quality records beyond the quality of the traditional architectural section drawing.

Google SketchUp
Holding to concerns of cost, ease of use, and availability, the Project has also investigated the best means for visualising our results, testing hypotheses in a 3D reconstruction environment and ultimately providing 3D reconstructions of our areas to the general public. In such work, a number of potential software packages present themselves for consideration, especially the well-known and often expensive professional 3D modelling systems such as 3DStudioMax, Maya, and even AutoCAD, to name a few. This Project has sought to work around expensive solutions, and has made great use of Blender3D, an open-source and cross-platform alternative to these packages. Since all of these methods present a relatively steep learning curve, however, the Project has also made considerable use of Google SketchUp, especially for the production of 'rough and ready' models for the testing of theories, the visualisation of spatial relationships, and the communciation of ideas within the Project. Though this modelling software does not permit the same level of detail as some other, it bears many advantages in cost, ease of use, flexibility, and compatibility which we have found to be useful. Indeed, even those with little familarity with 3D reconstruction can begin to experiment quickly and easily with the data and produce accurate, scaled, and thoroughly useful reconstructions.

Though SketchUp itself saves data in a closed format, it is a relatively simple task to export the 3D geometry produced into more easily transferable and archivable formats. Furthermore, the end product of reconstructions could utlimately be uploaded to the SketchUp Warehouse, and even used to populate Google Earth with data from the site. These web-deliverable advantages hold great potential for dissemination of ideas and the release of data to the general public. It is in this role that this software has been the most useful to the Project.



Vista outside the Porta Ercolano as painted by Niccolini (Le case ed I monumenti di Pompei Vol. II.) SketchUp Reconstruction of the area recorded by Niccolini
Website Content © Copyright Via Consolare Project 2014