Doerr, Martin and Apostolis Sarris (eds) 2003. The Digital Heritage of Archaeology. CAA2002. Computer Applications and Quantitative Methods in Archaeology. Proceedings of the 30th CAA Conference, Heraklion, Crete, April 2002. 295 Recognition and Classification of Fragments from Ceramic Artefacts Luigi De Napoli, M. Laura Luchi, Maurizio Muzzupappa and Sergio Rizzuti Università degli Studi della Calabria, Dipartimento di Meccanica, Ponte Pietro Bucci 44/C, 87030 Arcavacata di Rende (Cosenza), Italia {ldenapoli, luchi, muzzupappa, rizzuti}@unical.it Abstract. The paper describes a procedure for the analysis of fragments from ceramic artefacts The methodology employs the well known Reverse Engineering approach to suggest the collocation of a fragment on the surface of the vessel to which it seems to belong. The paper discusses the methodology and reports its validation, conducted by means of a numerical simulation and its application, for the recognition and classification of two real fragments belonging to two Apulian artefacts. The devel- oped procedure can be considered a valid aid for the archaeologist in order to speed up his/her activity. Keywords. Reverse Engineering, Surface Modeling, Surface Fitting, Curvature Analysis 1 Introduction This work reports the activities that have been carried out at the Department of Mechanical Engineering of the University of Calabria (Italy), for a PRIN project (Progetto di Ricerca di Rilevante Interesse Nazionale), in collaboration with the chair of “Methodology of Archaeological Research”, Bachelor de- gree in Tourism Sciences at the same University. The main task of the project consists of the development of a computer aided tool that aims to simplify and speed up the activity of an archaeologist, skilled in filing ceramic finds. The project is particularly concerned with finds (typically vessels) that are associated with mass manufacturing, when the number of fragments is high, and, in this context, one of the goals for the archaeologist consists of determining the total number of complete items that can be found in a site. The purpose of the project is to provide the archaeologist with a tool able to automate the counting of archaeological finds. This can be done only if the position of a fragment on a vessel can be recognized automatically. This basic aspect will be investigated in this paper. Some techniques, well known in engineering science, such as surface modeling and reverse engineering, have been em- ployed in order to create, in a virtual environment, the operat- ing conditions to evaluate the intrinsic geometric characteris- tics related to each fragment, that can be characterized by Gaussian and principal curvatures. The comparison of these estimated values with those computed on a virtual vessel al- lows the fragment to be located unambiguously. The paper reports the mathematical aspects related to the employed methodology and discusses the numerical simulation of the procedure. Furthermore, the recognition and classifica- tion of two real fragments is presented. 2 Brief Historical Background The work is related to vessel artefacts coming from a site located in the province of Cosenza in Northern Calabria. The place is called Cozzo la Torre, near Torano Castello, and the site is interesting from an archaeological point of view, be- cause in it settlements dating from the twelfth to the second century BC have been located (de La Genière, 1977). Fig. 1. Map of Cozzo la Torre, Calabria, Italy. The site is on a hill, on the west slope of the Crati valley (Fig. 1). It was a fortified settlement, populated in the fourth and third centuries BC by the Brettioi (or Brettii or Bruttii), an indigenous Calabrian people, who for many centuries had been influenced by Hellenic culture. The work has been car- ried out on fragments of two types of Apulian made artefacts (Lippolis, 1996), found in surveys (Genovese, 1990), (Geno- vese, 1999) (Fig. 2): - kraters, open shapes, typically with a diameter of 350 mm at the rim and 350 mm high; - skyphoi, open shapes, typically with a diameter of 130 mm at the rim and 130 mm high. The manufacture of all these types of pottery is very elabo- rate. In fact, most of these were decorative objects to show status.