Site Detection Using Radar Sentinel-1 and Optical Sentinel-2 Images Over the Thessalian Plain, Greece

This study explores the potential use of integrated radar and optical freely and open access distributed satellite images, namely the Sentinel-1 and Sentinel-2, respectively, to detect areas with archaeological interest. The Google Earth Engine platform was used to extract the backscattering signal from the Sentinel-1 sensor and the reflectance values from the spectral bands of Sentinel-2. A new image composite was produced over the Thessalian plain in central Greece and then imported into the ArcGIS Pro environment. The Support Vector Machine, Random Forest and Maximum Likelihood classifiers have been applied to identify areas that share similar radar and spectral profiles as other known sites. The overall findings indicate that despite the limited training sample used, the combined signal from both Sentinel sensors could detect up to 60% of the known sites.