Cultural landscapes from a bird’s eye view: applying aerial photography to identify archaeological sites in forest zone

The article shows the effectiveness of using aerial photography to detect and identify archaeological sites in forests. To test this method, we chose the well-studied archaeological complex of Shaitan in the south of Tomsk region for which detailed topographic plans and 3D elevation models are available, obtained via tacheometric survey on the ground. From the air, we examined eight sites: the hillforts of Shaitan I, II, III, and IV, the wider settlement of Shaitan I and the wider area of Shaitan I, as well as the burial sites of Shaitan II and III. The survey was performed by an unmanned aerial vehicle Geoscan 201, at the height of 330 meters, with the use of a digital full-frame camera Sony RX1 (35.8 mm x 23.9 mm, 6000 x 4000 pixels, 35 mm focal length), which allowed us to have images with resolution of about 5.6 cm per pixel. The photographs were taken with at least 50% of side overlap and 70% of forward overlap. The total area surveyed amounted to 120,000 m². Aerial photography data were processed using digital photogrammetry and Agisoft Meatshape software. As a result, we produced an orthophoto and a digital elevation model of high resolution which were analyzed in QGIS software. Viewing data in the hillshade layer we were able to identify large defensive constructions-ditches and ramparts present on the fortified sites. To search for smaller objects in the terrain such as minor depressions, we used the single colour gradient mode, with height specifically adjusted for each site. The UAV-collected data allowed us to have a detailed picture of terrain’s surface in the deciduous forest zone, excluding, however, the earth’s surface under coniferous trees. The data collected in the UAV-enabled survey in the deciduous zone were almost as accurate as the data collected in the survey on the ground (with an estimated error of up to 10 cm due to the presence of grass). During data processing, difficulties were encountered in aligning photographs for some areas beyond the territory under survey which featured homogeneous surface texture and no clearly distinguishable objects. To avoid such issues, future surveys in forest zones could introduce greater imagery overlap. Overall, we demonstrated the effectiveness of using aerial photography to search for archaeological sites and monuments in the forest zone, with a major limitation here being associated with coniferous forests. Aerial photography surveys conducted in deciduous forests before tree leaves appear and after they have fallen have proved to be giving excellent results.

Keywords

Authors

References