The problem is odder than that, though. The spikes project horizontally, not just vertically, and are much longer than any possible difference in height between cells. And looked at up close, they really are spikes projecting from an otherwise pretty intelligible circle. I think it must be related to relative sizes and resolutions: I'm using a 90 m DEM, draping a Landsat TM image at 30 m resolution onto that, but some of the circles are on the order of 100 m or smaller, and they are made up of 360 smaller vectors - so it's trying to draw very small vectors over a very coarse DEM. I haven't been able to figure out whether they're coming off where two cells meet in the DEM, or something like that. I've had no problems draping vectors over a 1-meter or 2-meter DEM. I really suspect this is a bug in NVIZ.
Thanks,
Nick Cahill
On Thursday, January 2, 2003, at 06:04 PM, H. Bowman wrote:
I am draping circles produced by v.circle and other GRASS programs over
a DEM in NVIZ, and getting odd results. Depending on the resolution of
the region when I start NVIZ, and the polygon size within NVIZ, I get
spike-like vectors projecting from the circles, either up and down,
north-south or east-west.The vector line of the circle will be drawn on the 3d surface. If there are lots of small hills and valleys in the DEM, the circle will appear jagged when not viewed from directly overhead.
What I've done for drawing vertically smooth coastlines is to make a new constant surface in NVIZ, of elevation 0 and transparency 255, and only display the vector file on that (flat) surface.
Maybe make a circle with r.circle and use it as a color mask for the DEM surface in NVIZ or some varation on that theme?
best of luck,
Hamish