[GRASS-user] Generalizing SRTM DEM for hillshading

[Lars Pennig]

Dear all,

I'm trying to generalize the 90m SRTM DEM for use in cartography, e.g. for a hillshading in a 1:700.000 map. Emphasis is on a good-looking map, not on a hydrologic correct model. 
For this scale, there is too much detail in the SRTM DEM, so I tried different filtering approaches: r.neigbors, r.resample.rst and other methods like resolution-bumping, but the output looks to soft: all the main ridges, which should give a quite crisp image have disappeared. I would like to keep the main structures (like high peaks and long ridges) and drop the small structures like short ridges, small gullys and channels.
I know, automated generalizing of elevation models is one of the biggest topics in modern cartography, but maybe someone has a good hint :-)

Thanks in advance, Lars