PubAg

Main content area

UAV vs classical aerial photogrammetry for archaeological studies

Author:
Nikolakopoulos, Konstantinos G., Soura, Konstantina, Koukouvelas, Ioannis K., Argyropoulos, Nikolaos G.
Source:
Journal of archaeological science: Reports 2017 v.14 pp. 758-773
ISSN:
2352-409X
Subject:
aerial photogrammetry, altitude, archaeology, computer vision, georeferencing, hills, models, orthophotography, surveys, unmanned aerial vehicles, Greece
Abstract:
Excavations at Stavros, Chalandritsa, 20km south of Patras, Greece, have brought to light an extensive Late Bronze Age settlement, inhabited at least during the 13th and 12thcenturies B.C. The settlement covers an area of approx. 12 acres on top of a rocky hilltop, which offers an unobstructed supervision of the wider region. As highly accurate mapping and 3D reconstructions are fundamental for analyses and interpretations in archaeology, the specific area was used as a test area for a comparison between classical topographic survey, airphoto and Unmanned Aerial Vehicle (UAV) based photogrammetry. The settlement was mapped with traditional topographic surveying methods using tachymeter and Differential GNSS system. Digital photogrammetric techniques were applied on analogue airphotos to create (DSMs) and orthophotos from the archaeological site. Two independent low altitude aerial campaigns were carried out using two different UAVs. Computer vision techniques along with photogrammetric analysis techniques were used to perform bundle adjustment with ground control points (GCPs) collected with a differential GNSS receiver. 3D models were created from the imagery captured from the UAV campaigns. Ultra-high resolution orthophotos with a pixel size of 5cm and DSMs with respective spatial resolution were also created. The orthophotos were validated in terms of georeferencing accuracy and the DSMs were validated in terms of height accuracy. The accuracy of the UAV derived products reaches 99.6% compared to classical topographic measurements. Chalandritsa test site results provide strong evidence that 3D models created by UAV imagery can be accurate enough to perform precise measurements.
Agid:
6126251