meshlab
meshlab copied to clipboard
cnr-isti-vclab
Faces decimation
on windows 10. Some years ago (same machine) I used often and successfully the items for "decimation" on large and complicated STL (scanned people). Now, even if I used a moderate decimation obtain many errors in the mesh. Even for very low decimation values. (not unexpected points but faces that form the third vertex in points belonging to the mesh, but not of adjacent faces.) Thank you! G.
The error mentioned in my previous message occurs mainly between coincident faces on the same plane. On the other hand, the problem is different between faces that form protruding angles to each other. The decimation does not take into account that some points are creating intersecting faces. In practice it could be that I do not know how to use it, even if I intervene with very light decimations. Example 30000 faces reduced to 27000, but I find those functions unusable. It may be that I never noticed because I was dealing with volumes attributable to "potatoes", but if the shape has a minimum of protrusions and indentations, even if I try to file a very limited number of faces, the resulting mesh is disastrous.
At this point I could select and delete the intersecting faces. But if I then ask to close the holes that form, the new closing faces create new intersections. I'm a little bit desperate ... help!
Thanks! The shape is many elaborate and I cannot publish it. But as soon as I have time I start over with a part and attach the file. There were two problems. The first is the intersections between faces, trying to slightly reduce the number of faces. The second is the closing of the holes caused by my automatic elimination of the faces that formed intersecated. I have to make a file because it is not easy for me to explain in English. These images are just to illustrate the type of work.
Ok, two comments: first you can try also to check the "planar simplification" option. Then, eventually, if you do not want to share the files publicly you can just send them privately ([email protected])
Many thanks, if I can isolate a small piece it is better, because it is easier to test, otherwise we do as you say!
Here, I have extrapolated a slice of my pine, which has no errors (chk.STL).
Then I attempt to apply "Quadric Edge Collapse Decimations", defaul setting, REDUCTION at 3000 faces (by original faces 3844).
Done....
Then I search for intersections.
Founded ... in picture.
chk.zip
Unfortunately I have no good news. I am preparing the basis of my model, it is very simple. The mesh is automatically obtained as the previous one by my scanning algorithm from 2d shading (clip).
The mesh is absolutely correct, neat and simple. (clip0)
Both menu reduction algorithms, applied to 50%, generate intersections between some faces ... (by clip1 to clip2)
I will try to extrapolate a critical part.
I tested the decimation by forcing the error with a small file that I am attaching. After loading it I set the quadric reduction to 7 faces. It seems that there is no control in the decimation algorithm in case intersections are generate
