The potential and implications of automated pre-processing of LiDAR-based digital elevation models for large-scale archaeological landscape analysis

0567036 - ARÚ 2023 RIV SK eng J - Journal Article
Novák, David - Pružinec, F. - Lieskovský, T.
The potential and implications of automated pre-processing of LiDAR-based digital elevation models for large-scale archaeological landscape analysis.
Slovak Journal of Civil Engineering. Roč. 30, č. 4 (2022), s. 1-10. ISSN 1210-3896
R&D Projects: GA MŠMT(CZ) EF16_013/0001439; GA MŠMT(CZ) LM2018134
Institutional support: RVO:67985912
Keywords : DEM filtering * landscape archaeology * GIS analysis * visibility * land surface curvature * drainage networks
OECD category: Archaeology
Impact factor: 0.4, year: 2022
Method of publishing: Open access
https://sciendo.com/article/10.2478/sjce-2022-0022

LiDAR-derived digital elevation models (DEMs) have transformed the archaeological study of landscape features, broadened our technical capabilities, and enhanced the accuracy with which terrain relief is described. These models also place demands on how researchers and analysts interpret DEM content in the context of the modern landscape. LiDAR-based DEMs contain modern man-made structures that can significantly influence model properties. Although data are usually filtered and some of these artificial features are removed during bare-earth classification, many terrain interventions remain visible. This large-scale case study applies established methods to a freely available DEM of the Czech Republic in an attempt to evaluate differences between original and filtered DEMs. It applies a fully automated filtering procedure using vector topographic maps to avoid manual corrections that would make the procedure problematic when used on a macro scale. The results of our archaeological GIS analysis demonstrate that this procedure, despite its relative simplicity, can achieve a significantly better representation of a landscape compared to that offered by an unfiltered DEM. Finally, we propose a series of future steps with a view to developing a more comprehensive and accurate model and overcoming its limitations.
Permanent Link: https://hdl.handle.net/11104/0338306


Research data: Zenodo