I am making shaded relief maps from USGS DRG scanned topo maps, and USGS DEM data, using GRASS 5.0beta10, compiled on Redhat 6.2.
The original horizontal resolution of the DEM is 30m. The resolution of the DRG is about 2.4m. So to produce the required smooth high resolution aspect file, I ran r.slope.aspect on the original DEM, then reset the region resolution and resampled the result to match the high DRG resolution, then ran r.neighbors on that file with method=average and size=15. This smoothed the hi-res aspect map nicely, except that near the edges of many dark, southeasterly sloping areas, there appeared many bright lines. They seem to follow along the edges of where the pixels were in the original. I have a nice snapshot of the effect which I could e-mail to anyone interested (188kB).
These lines are eleven pixels wide, with "fuzzy" edges, so they obviously have been averaged. However, there is no sign of discontinuities in the original aspect map. I ran r.neighbors again on the resampled file, with size=3, the same thing occurred, but the lines are fewer, and narrower.
I'm new to GRASS, so this is likely my mistake.
Am I doing something wrong, or is this a bug?
--
-----------------------------------------------------------
Carl Brown cbsled@ncia.net
-----------------------------------------------------------