Logo Roberval UTC

Understanding the daily lives of fossil humans involves interpreting the tools (stone, bone, antler, shell) of these past users. Archaeological research has been addressing these issues for many years. As with any object (and a fortiori stone tools, which will be the focus of the project presented here), use and exposure to different agents modify the topography of the surfaces most directly accessible to the environment. Thus, observing and characterizing the traces (topographical signatures) on the surface of tools can provide clues to the processes undergone: movements performed (scraping, sawing...), materials worked on (skin, wood...). To make the characterization and interpretation of stone tool surfaces alterations more robust, we propose to set up a method for multiscale characterization of the topography of altered surfaces using a corpus of experimental tools from different rocks (flint, quartzite, metarhyolite...) used (manually and with the aid of a tribometer) to saw or scrape different materials (bone, deer antler, reed...) with different numbers of cycles. This corpus will enable us to establish strategies for identifying and measuring areas of contact with the material being worked. Morphological indicators will then be set up to describe topographical signatures and thus move towards a classification of uses. The contributions of a classical statistical approach versus a deep learning approach will be assessed. The results will also be used to develop better excavation and post-excavation protocols to prevent alteration and better preserve traces of use on archaeological stone tools.

Plus d’informations :
[Télécharger l'annonce - PDF 110 Ko]