Open access
Author
Date
2017Type
- Doctoral Thesis
ETH Bibliography
yes
Altmetrics
Abstract
The main objective of this thesis is the reproduction of the style of rock depiction featured in the Swiss National Map by digital means. Since the production of such rock hachures is still done manually by trained specialists, automating the process is going to benefit experts and laypersons alike, who want to include Swiss style rock hachures into their maps. Additionally, automation offers fast, reproducible and economically feasible results.
After describing the anatomy of rock hachures, and a brief historical survey, current approaches to automating rock depiction are discussed, resulting in the following principles for the development of the novel method.
Within a map, rock depiction is no isolated concept. It is intimately connected to the depiction of shaded relief. Both map elements represent a lighting model that is focused on preserving contrast across terrain edges like ridges and ravines. This model is quite different from analytical Lambertian shading employed in many current approaches. On a structural level, shaded relief is simpler than rock hachures: a variation of grey values versus multiple kinds of jittered hachures, i. e. broad form strokes and narrow contour strokes for indicating the rock skeleton, as well as fill hachures depicting light-modulated faces. From this structural disparity follows the primacy of the shaded relief in the workflow. The rock hachures are derived from the shaded relief. As with all map elements, the issue of generalization needs to be addressed for rock hachures and the shaded relief.
These principles are implemented as follows. The user supplied raster elevation model is generalized along slope lines using line integral convolution. For the purpose of determining terrain edges, two shaded reliefs are generated, using a main lighting direction and an orthogonal direction. The terrain edges are computed by merging the discontinuities of the two shaded reliefs, extracted using an edge detection algorithm. Grey values are assigned to either side of every terrain edge, while care is being taken that the assignments maintain contrast and generally observe the main lighting direction. The grey values of the remaining faces between terrain edges are computed by solving a Laplace equation, where the already assigned grey values serve as the boundary condition. The rock faces are covered with fill hachures, where the stroke width is derived from the grey values of the shaded relief. The form and contour strokes are derived by offsetting the terrain edges according to the gradient of the shaded relief.
The method is automatic in the sense that, apart from the raster elevation model and a mask indicating the raster cells covered by rock, the user is only supposed to enter declarative parameters, e. g. the main lighting direction. It follows that in using this method, expert knowledge is no decisive factor. Different users providing the same input will arrive at equal results. Show more
Permanent link
https://doi.org/10.3929/ethz-b-000201368Publication status
publishedExternal links
Search print copy at ETH Library
Publisher
ETH ZurichSubject
Terrain generalization; Relief shading; Rock drawing; HATCHING + HACHURES + ROCK-DRAWING (CARTOGRAPHY); RELIEF MAPPING, TWO DIMENSIONAL (CARTOGRAPHY); CARTOGRAPHIC DRAWING TECHNIQUES; BERGSTRICHE + SCHRAFFEN + FELSZEICHNUNG (KARTOGRAPHIE); RELIEFKARTOGRAPHIE, ZWEIDIMENSIONAL (KARTOGRAPHIE); KARTOGRAPHISCHE ZEICHENVERFAHRENOrganisational unit
03466 - Hurni, Lorenz / Hurni, Lorenz
Related publications and datasets
Notes
Funded by SNF grant "Terrain analysis, feature extraction and model deformation for cartographic generalisation and visualisation" (200021-127224).More
Show all metadata
ETH Bibliography
yes
Altmetrics