I have being trying without success to find algorithm descriptions or source codes about bridging two triangular faces.... I would preciate any help.
triangle bridging
Author
Hi
I have being trying without success to find algorithm descriptions or source codes about bridging two triangular faces.... I would preciate any help.
I have being trying without success to find algorithm descriptions or source codes about bridging two triangular faces.... I would preciate any help.
Author
Ops! after several days without a response I'm fear there's no knowledge base about this in this forum? please at least suggest me another forum and I will leave forever this one and I will no bother you all with any more questions.
my last shot:
I need to bridge/connect two holes in mesh, I have several loop rings of vertex corresponding to each hole but all I found in internet are short verbal descriptions of succesful algorithm but no theory at all, someone has suggest me a modified Bresenham algorithm O.o but i'm in darkness.....
Thanks in advance
my last shot:
I need to bridge/connect two holes in mesh, I have several loop rings of vertex corresponding to each hole but all I found in internet are short verbal descriptions of succesful algorithm but no theory at all, someone has suggest me a modified Bresenham algorithm O.o but i'm in darkness.....
Thanks in advance
Sorry, but I think the problem is your description, I can't really understand what you want to do, can you provide some pictures?
Triangulation of holes?
Depending on what you want, this may interest you:
http://www.cs.cmu.edu/~quake/triangle.html
Triangulation of holes?
Depending on what you want, this may interest you:
http://www.cs.cmu.edu/~quake/triangle.html
Author
Oh thanks for your reply!!! you have bring me hope in this forum
here I attach you a picture of what I want to do, is not triangulating holes, is connecting holes in 3D space, what I need is a detailed algorithm descriptions and/or source code in order to understand it.
[attachment=5287:1.jpg]
[attachment=5288:2.jpg]
here I attach you a picture of what I want to do, is not triangulating holes, is connecting holes in 3D space, what I need is a detailed algorithm descriptions and/or source code in order to understand it.
[attachment=5287:1.jpg]
[attachment=5288:2.jpg]
It is still a form of triangulation though. Build a list of all vertices that are part of the holes, and triangulate the point cloud that they form (and you need to exclude from the final result any triangle that shares a edge with the original mesh).
Something like this
http://www.cse.unsw.edu.au/~lambert/java/3d/hull.html
Something like this
http://www.cse.unsw.edu.au/~lambert/java/3d/hull.html
Just a correction, "and you need to exclude from the final result any triangle that shares a edge with the original mesh", well, not true, should be
"and you need to exclude from the final result any triangle that shares more than one edge with the original mesh"
"and you need to exclude from the final result any triangle that shares more than one edge with the original mesh"
I think there's a more specific solution to this problem. Writing a robust point-cloud triangulation algorithm that handles non-convex geometry (non-convex holes) can be quite hard, and seems to be overkill for this specific problem.
I have to look into it though, I remember a whole topic about this problem in Descriptive Geometry (connecting holes for pipe-like structures), but that's in Hungarian, and I can't even remember the name of the topic. Tomorrow, I'll be back with some pointers, I guess
But it seems there can be many number of solutions though. Are there any constrains that suit your need, or any "decent looking" solution would suffice?
I have to look into it though, I remember a whole topic about this problem in Descriptive Geometry (connecting holes for pipe-like structures), but that's in Hungarian, and I can't even remember the name of the topic. Tomorrow, I'll be back with some pointers, I guess
But it seems there can be many number of solutions though. Are there any constrains that suit your need, or any "decent looking" solution would suffice?
Here's one idea:
-take the edge loop ("1st hole") that has the more vertices
-loop through all its points in order (march through the vertices of the edge loop)
--get the vertex from the other edge loop ("2nd hole") that's nearest to the current point. If it's the same as the stored vertex (see later) or if it's the "next vertex" after the stored vertex, then store its index. Else get another the second/third closest vertex. "Next vertex" means marching to the next point with consistent winding. For example in the 1st loop march through the vertices so that the triangles are on its right side, and march through the 2nd loop so that the triangles are on its left side.
I hope that's clear, I'm not sure about it at midnight with this English knowledge...
EDIT: Well, the closest point as the condition is not good, or at least not enough (it's very easy to come up with an example that would produce wrong result). somehow, the "curvature" of the curves have to be considered.
OR MAYBE: you get the two closest points on the two loops, THEN you start the proposed algorithm. Would this give correct results?
-take the edge loop ("1st hole") that has the more vertices
-loop through all its points in order (march through the vertices of the edge loop)
--get the vertex from the other edge loop ("2nd hole") that's nearest to the current point. If it's the same as the stored vertex (see later) or if it's the "next vertex" after the stored vertex, then store its index. Else get another the second/third closest vertex. "Next vertex" means marching to the next point with consistent winding. For example in the 1st loop march through the vertices so that the triangles are on its right side, and march through the 2nd loop so that the triangles are on its left side.
I hope that's clear, I'm not sure about it at midnight with this English knowledge...
EDIT: Well, the closest point as the condition is not good, or at least not enough (it's very easy to come up with an example that would produce wrong result). somehow, the "curvature" of the curves have to be considered.
OR MAYBE: you get the two closest points on the two loops, THEN you start the proposed algorithm. Would this give correct results?
OR MAYBE: you get the two closest points on the two loops, THEN you start the proposed algorithm. Would this give correct results?
Not enough....
The curvature cannot be discarded.(Shit, I hate it when others' computing problems are more interesting to me than my own...)
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement