BGR Bundesanstalt für Geowissenschaften und Rohstoffe

AGNES: Automated generalisation of geological spatial data

Country / Region: Germany

Begin of project: June 1, 2021

End of project: February 29, 2024

Status of project: December 1, 2022

Up to now, the geological general map series of the BGR were not data technically linked to each other. I.e. the GK1000 was previously derived "manually" from the GÜK200 or GÜK250.
Within the framework of the AGNES project, the FME-based generalisation tool AutoGen of the State Office for Geology, Raw Materials and Mining in Baden-Württemberg (Schuff 2019, LGRB) is being further developed or adapted on the basis of the above-mentioned two general maps, so that a largely automated derivation of small-scale geological BGR spatial data from large-scale geological BGR spatial data will be possible. In the future, this procedure should also be applied for other use cases, e.g. for the derivation of the German part of the International Quaternary Map of Europe 1:2,500,000 (IQuaME2500) from the GK1000.

Preliminary result of generalisation for the topic Genesis (left: GÜK250, right: preliminary GK1000)Preliminary result of generalisation for the topic Genesis (left: GÜK250, right: preliminary GK1000)

General functioning of AutoGen:
Before the actual start of the generalisation procedure, it is necessary to create a hierarchical mapping table for the semantic aggregation of the spatial data and a file with geometric parameters, both of which are integrated into the procedure. AutoGen consists of two vector-based and one raster-based generalisation stage. First, the geological units (areal vector data) are aggregated, typified and resolved using the previously entered semantic and geometric parameters and several generalisation steps. This is followed by a transformation of the newly created geological units from the vector data format into the raster data format and then a simplification/smoothing of these using a cellular automaton (CA). The CA calculates the requested final state at the specified target scale in an iterative process. Each further iteration leads to a recalculation of the state of a raster cell depending on the state of its neighbouring cells and thus to a simplification or generalisation of the raster data. Finally, a re-transformation into the vector data format (shape file) takes place.

Contact:

    
Dr. Marc Filip Wiechmann
Phone: +49-(0)511-643-3346
Fax: +49-(0)511-643-2304

This Page: