While the LiDAR-technology solved some problems of archaeological prospection and different visualisation techniques have been developed, the automated detection of cultural heritage still poses a significant challenge. Therefore geographers from the Ruhr-Universität Bochum and archaeologists from the Landschaftsverband Westfalen-Lippe in Münster are developing workflows for the automatic detection of diverse types of field monuments. In 2016, a Master’s thesis used object-based image analysis (OBIA, implemented in eCognition) for the detection of remains of burial mounds, motte-and-bailey castles and ridge and furrow field systems in a special terrain model named difference map, that was introduced by R. Hesse in 2010. The resulting areas of interest are available in shapefiles and manual interpretation is necessary, which still cannot and should not be done by a computer. However, this step is simplified because the results are classified by its condition. In this way, areas with probably well preserved field monuments are separated from others and can be interpreted at first without losing those that appear eroded. The results show, that all three types of field monuments are basically detectable with automated workflows and that the correctness can reach 75-90%. Furthermore, new field monuments were already found. After Celtic fields were found in Westphalia for the first time, research currently is expanded in order to develop a GIS-tool for calculating the required digital terrain models and searching for Celtic fields, by a template matching methodology integrated in Python. (M. Fabian Meyer, Ingo Pfeffer, Carsten Jürgens)
0 Comments