Abstract
The Ludwig Boltzmann Institute for Archaeological Prospection and Virtual Archaeology (LBI ArchPro) together with its partner the Central Institute for Meteorology and Geodynamics (ZAMG) has developed motorized magnetic prospection systems to survey entire archaeological landscapes within reasonable time at high spatial sampling resolution. With these novel prospection system, the LBI ArchPro has successfully surveyed in total more than 32 square kilometres over the past five years resulting in magnetic maps containing millions of anomalies. While data collection data positioning and the processing of the data have been automated (with exception of the operation of the survey vehicles), the outlining (data segmentation and classification) and interpretation of the prospected magnetic anomalies has become in regard of large amount of generated data a tedious, very time consuming, so far manually conducted task. In order to speed up the interpretation process we have therefore developed a workflow and algorithms for the automatic detection, outlining and classification of magnetic anomalies. Relevant magnetic prospection anomalies are automatically classified into two classes, which are "iron litter'' objects that are located dosed co the surface, and deeper reaching individual "pit" objects. Based on this classification we calculate several physical and geometrical properties for each object and export this data to a Geographical Information System (GIS) for further interactive classification and subsequent data interpretation.