SCIentific RESearch and Information Technology Ricerca Scientifica e Tecnologie dell'Informazione Vol 12, Issue 2 (2022), 87-98 e-ISSN 2239-4303, DOI 10.2423/i22394303v12n2p87 Open access article licensed under CC-BY-NC-ND CASPUR-CIBER Publishing, http://www.sciresit.it MULTISPECTRAL IMAGING AND ARTIFICIAL INTELLIGENCE FOR ARCHAEOLOGY: FIRST RESULTS AND FIRST PROJECTS Cyrille Chaidron*, Sébastien Lermenier** *Université de Picardie Jules Vernes (UPJV) – Amiens, France; ArScAn-GAMA -CNRS, Paris; CEO of ARTEKA **CTO of Arteka Abstract Archaeology is a human science that is very attentive to technological evolution. For decades, it has relied on technologies from the hard sciences (medicine, chemistry, geology, etc.). The emergence of satellite images with increasingly fine resolutions and the massive arrival of the drone in the field of archaeology have created new uses for the detection of archaeological sites. Multispectral imagery now supports other technologies (geophysics, Lidar). But it is above all the arrival of artificial intelligence and the development of Deep-Learning that is taking archaeology into a new era. The large amount of documentation generated by archaeology is conducive to the development of projects that will use artificial intelligence to help archaeologists in their research and enable them to obtain new results, both in the detection of archaeological sites and in the analysis of artefacts such as ceramics. Keywords Archaeology, multispectral imaging, satellite images, artificial intelligence, ceramology, roman pottery 1. Introduction Archaeology is a discipline of the human sciences which very early on had to rely on technological developments in order to be more efficient as soon as the pressure of real estate development became apparent. Archaeologists, in the context of preventive archaeology (i.e. archaeology that takes place before development work is carried out), have had to reinvent themselves in order to meet legislative requirements and economic pressures. This was achieved firstly through technical improvements in the "field" phase (the archaeological excavation as it is understood, i.e. when archaeologists are on the dig). Mechanisation has become more and more important. Excavation techniques and recording methods have also been developed and improved, but it is during the study phase, i.e. after the excavation (the so-called 'post-excavation' phase) that archaeologists have to 'dip into' the hard sciences to acquire innovative methods of analysis. Archaeologists have drawn on technologies from the space field (Lidar, for example), from the field of medical imaging (X-rays for corroded objects) and from the field of geology (petrography for pottery fragments). Archaeology has also seized upon recent technologies for the general public such as VR and AR in the context of the enhancement of archaeological excavations and exhibitions. The multidisciplinary nature of archaeology is a breeding ground for innovative experiments, which explains why it is certainly the most “technology enthusiast” discipline in the humanities and social sciences. It is therefore not surprising that the innovative technologies of multispectral imaging or that of artificial intelligence (or augmented intelligence, dear to the great specialist Luc Julia) find a favourable ground for their development, even if it is still timid. This is what we are talking about here through some projects carried out within the Trame laboratory of the University of Picardy and within the start-up Arteka.