Author(s): Markus Bahner; Yannic Fuchs; Stefan Schafer

Linked Author(s): Yannic Fuchs, Stefan Schäfer

Keywords: Image processing; Feature extraction; Breaklines; Mesh refinement; Bedload transport

Abstract: With rapid development of measurement technology and geodata collection, the amount of data to be handled for hydraulic investigations is increasing. This study explores the potential to process data from a variety of rasterized sources with an image processing method based on Cannys algorithm. We present a processing workflow and its application in four different case studies. In the first case, a high-resolution manual survey of an alpine creeks active channel is handled. Herein, the program extracts structural elements and contours of the embankment in their 3D vectorized form, which were used to locally increase the mesh node density of a CFD-model. The second dataset is an ALS-based DTM of the floodplain of the Schmutter River in Bavaria. Here, the method is used to extract breaklines of levees, roads, and railways, in order to optimize the representation of the floodplain. The third case explores a multibeam echosounder bathymetry survey of the Inn River. The particular focus is the automatic measurement of the size of dunes in the riverbed. The final case deals with plane view photographs of a laboratory experiment concerning sediment transportation. Like in the third case, the measurement of the size of transport bodies is attempted with the presented technique. The successful implementation of the developed image processing method in the different cases demonstrates its versatility of use. The technique can be used to pre-process meshes for CFD models, as well as to improve post-processing methods of lab and field data. Overall, the results suggest that Cannys edge detector is a versatile and effective method to process data in hydraulic engineering.

DOI: https://doi.org/10.3850/IAHR-39WC252171192022735

Year: 2022