One of the benefits of NURBS curve over, say Bezier curve, is the ability to create offset curves exactly. How to proceed with such computation? Do I just translate the control points?
Asked
Active
Viewed 539 times
6
-
What do you mean exactly by offset curves? – Alan Wolfe Aug 25 '15 at 14:44
-
An "offset" curve has constant distance to the given curve, aka parallel curve. – Ecir Hana Aug 25 '15 at 14:47
-
I wonder what property of NURBS make them able to do this. I would think it would be that they are rational, but then rational bezier curves would also have this property. – Alan Wolfe Aug 28 '15 at 04:07
-
Where did you see this claim? It is rather surprising to me (but I don't immediately see that it's impossible). – Etienne Sep 16 '15 at 06:19
1 Answers
1
Quickly Googling produces:
- This related question on StackOverflow
- "Computing offsets of NURBS curves and surfaces" (paper)
- Discussions of 2.
To summarize them:
- Suggests that this is impossible to do exactly.
- Gives an algorithm to compute an offset curve approximately (though to within a tiny epsilon). If you can't access the paper, then:
- Look at any of these. This presentation for example seems to cover it.
geometrian
- 1,980
- 13
- 26
-
Bit of nitpicking its not impossible to do exactly. Its just not possible to do exactly for all possible cases. There are lots of cases where this can in fact be done exactly its just a bit hard to generalize this. – joojaa Sep 23 '15 at 23:40
-
@joojaa IMO possible cases are exceptions rather than the rule. Can you link to a reference ? – Sep 23 '15 at 23:50
-
Yes but thats enough to refute impossible. Not in front of a desktop right now bit consider this: Since rational splines can do exact circular arcs that means you can do exact offsets of said arc. Since linear splines are possible and their offsets contains lines and circular arcs that too is possible. A big set of realworld uses there. – joojaa Sep 23 '15 at 23:55
-
So if you say impossible to do exactly in the general case then its ok. – joojaa Sep 24 '15 at 00:04
-
Yes; I meant impossible in the general case (at least for Bézier). If you actually read the answers I was linking to when I said this, you'll find it says exactly that. – geometrian Sep 24 '15 at 02:16
-
1Point being, what you say should be self contained. You should update your post so that others who dont have the time dont make wrong conclusions. Remember discussions under posts are not permanent. What you dont say on the otherhand you can leave to the link. – joojaa Sep 24 '15 at 06:58
-