Application of Classification Methods on Geophysical Data from the Archaeological Site of Aptera, Chania, Greece

Classification techniques such as the self-organizing-maps (SOM) network are proposed as a tool for efficient interpretation of geophysical data obtained from archaeological investigation. This methodology was useful for the interpretation of georadar data from the archaeological site of Aptera, Chania, Greece where an integrated geophysical survey covered 9730 m2. The survey was realized under the auspices of the 25th Ephorate of Prehistoric and Classical Antiquities, Ministry of Culture. Soil resistance and magnetic gradient maps, resistivity and georadar sections and 3D images, aerial and ground photos, archaeological and geologic maps were registered in a GIS project. The geophysical maps display the ancient city plan. Resistivity depth slices exhibit a stone corridor at a depth of 0.34 m. Three-dimensional inversion was applied on electrical tomography data collected along parallel survey lines. The resulted three-dimensional image reveals a water pipeline. Instantaneous and geometric attributes calculated from the migrated georadar sections over a roman cistern were subsequently classified using SOM. The resulted classified 3D georadar image displays the arc-shaped roof of the cistern as well as stone pipelines which supplied the cistern with rain water.