Multispectral drone images analysis is the second new methodologies that will be tested during the phd project. Also in this case the goal of the project is testing the potential of a new methodology of non-invasive archaeological investigation and identifing elements that fill the empty spaces of the medieval landscape to reconstruct the archaeological continuum.
In the 1980s starts the multispectral images analysis did with archaeological purpose. In this very early phase the images were obtained from the MSS (multispectral scanner) and TM (Thematic Mapper) sensors mounted on the Landsat satellite constellation, having a resolution of 60 and 30m per pixel. In the late 90s, while the launch of the Landsat 8 brought the ETM + sensor into orbit (with resolution up to 15m in the panchromatic band), Sever was the first researcher to use the spectral indices to process multispectral images and detect large archaeological elements not visible on the surface. More recently, technological developments have made it possible to constantly improve the resolution of the sensors applied to satellites and aircraft (below one meter per pixel), and to apply new multispectral sensors to commercial drones, significantly reducing costs and times of research, and enormously improve the resolution of multispectral images (also thanks to the possibility of having images taken at a much lower altitude), opening up new developments in the field of archaeological research, with new potential still to be explored.
Unlike thermal sensors, which analyze the soil by detecting temperature differences, multispectral sensors investigate the soil indirectly: by analyzing the state of health of the vegetation that covers it. In particular, the vegetation reflects and absorbs the electromagnetic radiation in a different way, based on different physiological properties determined by the soil on which they insist, and in bands of the electromagnetic spectrum not visible to the human eye. For example, an healthy plant reflects sunlight more in the medium infrared (MIR = 2.5 - 20 μm) and near infrared (NIR = 0.7 - 2.5 μm) compared to an unhealthy one.
By analyzing the different state of health of the plants, it is possible to identify archaeological elements not visible on the surface. For example, a ground that cover a wall will result in less healthy plants, while plants that insist on a ground covering a moat will be healthier (see fig. 1)
Using the aerial perspective of the images taken by drone it will therefore be possible to highlight the alignments of more or less healthy plants, allowing to reconstruct the shape and size of the archaeological elements not visible on the surface (see fig. 2).
Agudo, P.U., Pajas, J.A., Pérez-Cabello, F., Redón, J.V., Lebrón, B.E., 2018, The Potential of Drones and Sensors to Enhance Detection of Archaeological Cropmarks: A Comparative Study Between Multi-Spectral and Thermal Imagery. Drones 2018, 2, 29
Fiorini L., Materazzi F., 2017, Un Iseion a Gravisca? Fotogrammetria, telerilevamento multispettrale da APR e dati archeologici per una possibile identificazione. FOLD&R FastiOnLine documents & research (396). pp. 1-23. ISSN 1828-3179
Verhoeven, G., 2008. Imaging the Invisible. Using modified Digital Still Cameras for Straightforward and Low-Cost Archaeological Near-InfraRed Photography. Journal of Archaeological Science 35(12), 3087-3100
Webber, Henry & Heyd, Volker & Horton, Mark & Bell, Martin & Matthews, Wendy & Chadburn, Amanda, 2017. Precision farming and archaeology. Archaeological and Anthropological Sciences. 10.1007/s12520-017-0564-8.