Normal inconsistency for the mesh generated by greedy projection triangulation tutorial

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Normal inconsistency for the mesh generated by greedy projection triangulation tutorial

jogging
Hi,all

I run the greedy projection triangulation tutorial using PCL1.6. The souce code and the cmake file is from
the link below:
http://pointclouds.org/documentation/tutorials/greedy_projection.php#greedy-triangulation.
After the mesh is generated, I save the mesh as a stl file using the function loadPolygonFileSTL.
I open the mesh in MeshLab. I think the mesh is not normal consistent because some triangles are bright and some triangles are dark.
Is this a problem?

Regards
Jogging
Reply | Threaded
Open this post in threaded view
|

Re: Normal inconsistency for the mesh generated by greedy projection triangulation tutorial

aichim
Administrator
Hi,

Are you sure it is the problem with normals or the dark triangles simply don't exist? Do you see them if you look at them from the other side?

This algorithm seems to have always been problematic and it seems that fine tuning the parameters and improving the quality of the input data (uniform density of the point cloud, consistent normals etc.) might get you better results.

Cheers,
Alex
---
http://www.alexichim.com
http://www.openperception.org




On May 15, 2013, at 5:13 PM, jogging <[hidden email]> wrote:

> Hi,all
>
> I run the greedy projection triangulation tutorial using PCL1.6. The souce
> code and the cmake file is from
> the link below:
> http://pointclouds.org/documentation/tutorials/greedy_projection.php#greedy-triangulation.
> After the mesh is generated, I save the mesh as a stl file using the
> function loadPolygonFileSTL.
> I open the mesh in MeshLab. I think the mesh is not normal consistent
> because some triangles are bright and some triangles are dark.
> Is this a problem?
>
> Regards
> Jogging
>
>
>
> --
> View this message in context: http://www.pcl-users.org/Normal-inconsistency-for-the-mesh-generated-by-greedy-projection-triangulation-tutorial-tp4027835.html
> Sent from the Point Cloud Library (PCL) Users mailing list mailing list archive at Nabble.com.
> _______________________________________________
> [hidden email] / http://pointclouds.org
> http://pointclouds.org/mailman/listinfo/pcl-users

_______________________________________________
[hidden email] / http://pointclouds.org
http://pointclouds.org/mailman/listinfo/pcl-users
Reply | Threaded
Open this post in threaded view
|

Re: Normal inconsistency for the mesh generated by greedy projection triangulation tutorial

jogging
Thanks for your response.

I am using the same point cloud as the example.



Even if I set the flag to false using   gp3.setNormalConsistency(true),
the problem remains.

I am reading the paper about greedy projection algorithm. The paper says
the result has variable density. But the density in the result is almost uniform.
Are the parameters very critical to use the function successfully?

Regards
Jogging

aichim wrote
Hi,

Are you sure it is the problem with normals or the dark triangles simply don't exist? Do you see them if you look at them from the other side?

This algorithm seems to have always been problematic and it seems that fine tuning the parameters and improving the quality of the input data (uniform density of the point cloud, consistent normals etc.) might get you better results.

Cheers,
Alex
---
http://www.alexichim.com
http://www.openperception.org




On May 15, 2013, at 5:13 PM, jogging <[hidden email]> wrote:

> Hi,all
>
> I run the greedy projection triangulation tutorial using PCL1.6. The souce
> code and the cmake file is from
> the link below:
> http://pointclouds.org/documentation/tutorials/greedy_projection.php#greedy-triangulation.
> After the mesh is generated, I save the mesh as a stl file using the
> function loadPolygonFileSTL.
> I open the mesh in MeshLab. I think the mesh is not normal consistent
> because some triangles are bright and some triangles are dark.
> Is this a problem?
>
> Regards
> Jogging
>
>
>
> --
> View this message in context: http://www.pcl-users.org/Normal-inconsistency-for-the-mesh-generated-by-greedy-projection-triangulation-tutorial-tp4027835.html
> Sent from the Point Cloud Library (PCL) Users mailing list mailing list archive at Nabble.com.
> _______________________________________________
> [hidden email] / http://pointclouds.org
> http://pointclouds.org/mailman/listinfo/pcl-users

_______________________________________________
[hidden email] / http://pointclouds.org
http://pointclouds.org/mailman/listinfo/pcl-users
Reply | Threaded
Open this post in threaded view
|

Re: Normal inconsistency for the mesh generated by greedy projection triangulation tutorial

jogging
Hi, all

I know why the normal of the mesh is not consistent. One flag can control whether the vertex order
of the triangles need to be consistent. One can set the flag by calling the function setConsistentVertexOrdering.

But even I set the the flag. The normals are not consistent exactly.


From this mailing list, Alexandru Ichim said in one post that NormalEstimation does not progogate
the normal directions along the point cloud. Is this the reason?

Several users complained that greedy projection algorithm will produce holes in the final mesh.
From your experience, which method is suitable for the merged points from different viewpoints?

Thanks
Jogging
Reply | Threaded
Open this post in threaded view
|

Re: Normal inconsistency for the mesh generated by greedy projection triangulation tutorial

rbaleksandar
Did you find a solution for this? I have the very same issue. And yes, the greedy triangulation produces holes in the mesh. The black triangles are indeed black (opened it in Meshlab too just to make sure that's the case).