A common problem in art conservation is the reconstruction of fragmented objects. The fragments need to be assembled to produce the overall original shape of the object. This process includes the estimation of their relative positions (fragmentation problem), and the testing of their fractured surfaces for whether they match. These problems become more difficult when a great number of large and heavy fragments (or small and fragile ones) need to be physically manipulated in order to test potential matches. This increases the risk of damaging the objects while performing unsuccessful tests. So far, the process of selecting candidate matches has relied on access to the real fragments and/ or photographic and manual records. The success of the process is dependent on the experience of the conservator and the quality of the records. Here, a new methodology for solving this problem is proposed using digital tools, which introduce minimal risk to the object. Modern photogrammetric and laser scanning equipment can be used for the remote accurate recording of objects. These methods produce groups of 3D points collected from the surface of the object (called point clouds). The point clouds are triangulated to form a 3D model, from which the outward surface normals for each triangle can be computed. These normals are then analysed in Riemannian space (spherical surface) by calculating the Mahalanobis distance for each normal. This analysis allows the division of the models into their facets (the procedure is called segmentation and the facets identified are called segments) by setting a fixed maxirnum distance value, as the criterion for normals belonging to the same segment. The identification of a pair of potentially matching segments, can be achieved using a combination of parameters including conservation-related criteria and quantitative data from surface analysis. The conservation-related criteria originate from the visual study of objects and models and are related to the structural and decorative features of the objects. The surface analysis parameters come from the calculation of the eigenvalues of a normals' dataset, which allows the characterisation of the isotropy, spread and size of the surface. The results of the analysis are plotted in a 3D chart and are combined with the conservation-related criteria in a specialised database. In this document, several case studies of the application of this methodology to fragmented archaeological objects are presented. These case studies show the advantages of the proposed recording techniques against the traditional ways of recording. They also demonstrate how it is possible to segment a model using the proposed analysis, to determine whether a surface is a fracture or not, and to identify the style of carving of an object by analysing the tool-marks. Most importantly, the results of the case studies demonstrate the vast potential of the method in identifying matching fragments from a group, without any physical contact with the object. The risk introduced in the procedure is minimal and the reconstruction takes place after all the original locations of the fragments have been found using the digital models.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:565937 |
| Date | January 2002 |
| Creators | Velios, Athanasios |
| Publisher | Royal College of Art |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.0024 seconds