Page 1 of 1
Surface Reconstruction Blob ?
Posted: Sun Jan 26, 2025 3:26 pm
by Marvin_Martian
I'm computing the normal's for the mesh below using Triangulation setting. As well I'm using Poisson Surface Reconstruction although this is the results, why does the mesh look like it's a blob, are my Reconstruction settings too high or low ?
- CloudCompare_6OWnvj74u1.png (465.84 KiB) Viewed 526 times
How can I tell Surface Reconstruction which side of the mesh to Reconstruct as I'd like to do only half ?
Re: Surface Reconstruction Blob ?
Posted: Mon Jan 27, 2025 10:12 am
by daniel
So with Poisson, you generally need to filter the result based on the 'Density' scalar field (see the last section of
https://www.cloudcompare.org/doc/wiki/i ... n_(plugin)).
You can aslo play with the 'boundary' setting to change the behavior of Poisson where no point is present.
Re: Surface Reconstruction Blob ?
Posted: Mon Jan 27, 2025 1:07 pm
by Marvin_Martian
daniel wrote: ↑Mon Jan 27, 2025 10:12 am
So with Poisson, you generally need to filter the result based on the 'Density' scalar field (see the last section of
https://www.cloudcompare.org/doc/wiki/i ... n_(plugin)).
You can aslo play with the 'boundary' setting to change the behavior of Poisson where no point is present.
- CloudCompare_LiDnn8Ns9f.png (587.64 KiB) Viewed 448 times
I understand although, using filter I assume I want to eliminate the areas where the yellow line points a but when doing that using filter or the display ranges slider it removes part of the mesh beyond the white line which is areas I want to keep.
Essentially, where the yellow lines are I want to remove to make a new mesh but also remove parts of the mesh where the white arrow points but not all the way towards the white arrow.
Re: Surface Reconstruction Blob ?
Posted: Mon Jan 27, 2025 5:10 pm
by daniel
Ah, so you are happy with some of the 'bubbles' :-D
In this case, it's tricky... Because the only information we have is how far the triangles are from the closest points. What you could do is sample a lot of points on the raw mesh, segment the resulting cloud with the scissors tool, and eventually compute a new mesh based on this segmented cloud...
On the sampled cloud, you might also be able to segment the points based on their normal orientation, with the Sterogram tool of the Facets plugin (
https://www.cloudcompare.org/doc/wiki/i ... am_display). But I don't know how it will work with such a complex shape.
Re: Surface Reconstruction Blob ?
Posted: Tue Jan 28, 2025 1:33 am
by Marvin_Martian
daniel wrote: ↑Mon Jan 27, 2025 5:10 pm
Ah, so you are happy with some of the 'bubbles' :-D
In this case, it's tricky... Because the only information we have is how far the triangles are from the closest points. What you could do is sample a lot of points on the raw mesh, segment the resulting cloud with the scissors tool, and eventually compute a new mesh based on this segmented cloud...
On the sampled cloud, you might also be able to segment the points based on their normal orientation, with the Sterogram tool of the Facets plugin (
https://www.cloudcompare.org/doc/wiki/i ... am_display). But I don't know how it will work with such a complex shape.
The scissor tool ?
I tried the stereogram_display method in the facets plugin, after applying normals then converting them to HSV. That tool is difficult to use it doesn't select just one half of the mesh, it select the top or the side but not just strictly one side of the mesh. Along with trying the stereogram tool, I tried the facets plugin. I see that it makes a ton of faces based on the point cloud but how do you merge the faces ?
Re: Surface Reconstruction Blob ?
Posted: Thu Jan 30, 2025 8:55 pm
by daniel
Ah ah no, forget about Facets, it was really just the Stereogram tool that allows to filter points based on their normals orientation (but yes, all points at once, with all normals pointing in a particular direction).
And I was referring to this tool:
https://www.cloudcompare.org/doc/wiki/i ... ation_Tool