Uavs for the Documentation of Archaeological Excavations

UAV photogrammetry experienced a growing variety of diverse applications in different scientific disciplines. Comparable early, UAVs were deployed in the Cultural Heritage and Archaeology domains, mainly for the purpose of monument, building and landscape modelling. In this paper, we will focus on the investigation of UAV application for documenting archaeological excavations. Due to the fact, that excavations are dynamic processes and therefore the objects to be acquired change significantly within few hours, UAVs can provide a suitable alternative to traditional measurement methods such as measuring tape and tachymeter in some cases. Nevertheless, the image processing steps have to be automated to a large amount as results, usually sketches, maps, orthophotos and 3D models, should be available temporally close to the related excavation event. In order to accelerate the processing workflow, an interface between the UAV ground control software and various photogrammetric software packages was developed at ETH Zurich which allows for an efficient management and transfer of orientation, trajectory and sensor data for fast project setup. The capabilities of photogrammetric methods using UAVs as a sensor platform will be demonstrated in 3 case studies: The documentation of a large archaeological site in Bhutan, an excavation of a smaller site containing ancient tombs which include several uncovered objects in the Nasca region in Peru and the Maya site of Copán in Honduras. The first and the third case study deal with the 3D modelling of buildings and their remains by means of photogrammetry, which means that accurate flight planning had to be applied and followed during the flights. In the second case study, we acquired various aerial images over the excavation area Pernil Alto near Palpa in a more simple way for quick documentation of the area of interest. In a third part, we will present our results from comparisons between the planned positions for image acquisition and the positions realized by the navigation unit during the flight for both UAV systems mentioned above. We will describe how accurate orientation data improve automated image processing if they are at hand directly after the flight and explain the workflow developed at ETH Zurich.