Maxwell Render

Hi
since there are so many rhino experts around, I dare to ask a not maxwell related questions:
what is the easiest way to create a terrain model in rhino?
I got a bunch of contour lines on the correct height and i'm not able to create an accurate model. It doesn't have to look smooth.
What I tried:
Lofting the lines -> doesn't work at all
manually creating an edgesurface between two lines works fine generally, but in some parts it produces weird intersections.
There'd also be the possibility to create a stepped model, but it lacks of accuracy.

any suggestions? I'd be glad if somebody knows a good solution.

thanks, kami

I use the pointset reconstruction tools (Delauny) from the Rhino Labs.

You could also build the contours like you would for a physical model made from cardboard, extruding the curves (solids) the correct amount (1' for 1' contours for example) then drape a mesh over them.

you could also try the patch command.
when you use high uv count, the surface will fit quite well.
If you don't need 100% fit to the given input, the advantage is that you get a smooth nurbs surface

_________________
workstation: EVGA SR-2 / dual x5650@3.5Ghz / quadro 4000

thanks for the suggestions.
I completely forgot about the drape tool, which is a nice and fast tool to get a quick smooth model (since I already got a cardboard-style-model). But it's not very accurate, like the patch tool (even with the highest u/v count of 50/50).
the pointset reconstruction tools look pretty nice, but I'm getting strange results on the file I'm working now
(two different settings of samples)
It might be a problem of the curves who originated from an imported dxf of the geologist. I extended those lines with polylines to get more surroundings. Is there any way to solve this, or is my approach completely wrong?
My intention is to model the existing landscape to create a garden. But there are a lot of restrictions by the law of how much terrain you are aloud to dig out or to fill up. For that reason the model has to be more or less accurate. My cardboard style model is not sufficient enough.

Thanks for every help!
kami

i used the point set plugin once or twice before... it also accepts points only input or? i used the Delaunay mode....
maybe it helps to extract all points from the polylines, and give it only point information as input, that should result in a homogenous mesh as long as points are not "overlapping in Z".

or if you have a "cardboardmodel" (i guess a "höhenschichtenmodell") you could project an array of points in top view back onto the cardboardmodel, to get a uniform triangulation when using Delaunay

_________________
workstation: EVGA SR-2 / dual x5650@3.5Ghz / quadro 4000

it takes curves too--your images look like there's a problem with the input curves/points.

have you tried solving it in pieces instead of trying to solve the entire mesh in one go?

my first try was with curves, which doesn't work too good. even if you select only a few curves.
my second try with the extracted points (thx for the tip, polynurb) works better, but still not very convincing.
I don't know if there exists a simple solution or if it needs a lot of tweaking to get a good result. I sadly don't have much more time for tryouts so I'll give the cardboard model another effort and maybe add some steps inbetween to get more precision.
I've uploaded the rhino file though, if somebody would like to play around with it:
http://deiters.ch/maxwell/terrain_lines.zip

But I hope I can use the pointset tools more successfully on another project in the future. Thank you very much for the help.
cheers, kami

kami, i looked at your file and i think there is no "in one go" solution.
I faced the exact problem a couple of times before...
(there is an imho overpriced "rhinoTerrain" plug in ( http://www.rhinoterrain.com/index.php ), which has more features than PtsetRec. I had a go with the trail version... its fine for vast landscapes.. in my case i ended up building all by hand.. took longer but i really got what i needed )

in your model there are steep walls and zones of high detail but also long continous slopes... i would try to split the existing curves along breaklines of the terrain , and seperate vertical from horizontal parts. then build the surfaces between the breaks part by part, patching lofting...etc.. also delaunay will work better with less(divers) input.

If you want to add more curves between the existing z slices, try _meancurve, it interpolates two curve into a new one.

_________________
workstation: EVGA SR-2 / dual x5650@3.5Ghz / quadro 4000

thank you polynurb. terrain is a very tricky part of modelling. I've always had problems with all other programs I used so far. I just hoped, Rhino (I've only been working for one year with rhino now) would have an "all in one solution". It's a pity that there isn't but maybe it's arithmetically impossible. But the approach you described should provide a usable terrain in a reasonable amount of time. I'll try it the next time.
Thanks also for the _meancurve command which works great (for most of the curves only )
cheers, kami

Not that I'll be making terrain anytime soon. But if I was, I would use
SoftMove .. go to the manual, go to Transforms, then Set Points, watch
the videos for SoftMove, there's 5 or 6 different types.

Built-In..

- np