Applying automated object detection in archaeological practice: A case study from the southern Netherlands
Abstract Within archaeological prospection, Deep Learning algorithms are developed to detect objects within large remotely sensed datasets. These approaches are generally tested in an (ideal) experimental setting but have not been applied in different contexts or ‘in the wild’, that is, incorporated in archaeological prospection. This research explores the applicability, knowledge discovery—on both a quantitative and qualitative level—and efficiency gain resulting from employing an automated detection tool called WODAN within (Dutch) archaeological practice. WODAN has been used to detect barro... Mehr ...
Verfasser: | |
---|---|
Dokumenttyp: | Artikel |
Erscheinungsdatum: | 2021 |
Reihe/Periodikum: | Archaeological Prospection ; volume 29, issue 1, page 15-31 ; ISSN 1075-2196 1099-0763 |
Verlag/Hrsg.: |
Wiley
|
Schlagwörter: | Archeology / History |
Sprache: | Englisch |
Permalink: | https://search.fid-benelux.de/Record/base-26850881 |
Datenquelle: | BASE; Originalkatalog |
Powered By: | BASE |
Link(s) : | http://dx.doi.org/10.1002/arp.1833 |
Abstract Within archaeological prospection, Deep Learning algorithms are developed to detect objects within large remotely sensed datasets. These approaches are generally tested in an (ideal) experimental setting but have not been applied in different contexts or ‘in the wild’, that is, incorporated in archaeological prospection. This research explores the applicability, knowledge discovery—on both a quantitative and qualitative level—and efficiency gain resulting from employing an automated detection tool called WODAN within (Dutch) archaeological practice. WODAN has been used to detect barrows and Celtic fields in LiDAR data from the Dutch Midden‐Limburg area, which differs in archaeology, geo‐(morpho)logy and land‐use from the Veluwe in which it was developed. The results show that WODAN was able to detect potential barrows and Celtic fields, including previously unknown examples, and provided information about the structuring of the landscape in the past. Based on the results, combined human‐computer strategies are argued, in which automated detection has a complementary, rather than a substitute role, to manual analysis. This can offset the inherent biases in manual analysis and deal with the problem that current automated detection methods only detect objects similar to the pre‐defined target class(es). The incorporation of automated detection into archaeological prospection, in which the results of automated detection are used to highlight areas of interest and to enhance and add detail to existing archaeological predictive maps, seems logical and feasible.