3D reconstruction

Hi, is it possible to do a 3D reconstruction in paraview?
I want to obtain a 3D mesh from this.
Do you have any suggestion?

I want something like this

Filters → Delaunay3D ?

Yes, I tried but it is like this:

Should I change some settings?

I’m afraid delaunay is limited for reconstruction, maybe @Charles_Gueunet has some inputs

1 Like

You can try increasing the Alpha parameter and enabling Alpha Lines and Alpha Verts but as Mathieu suggested the filter is limited for the type of reconstruction you are interested.

If you are familiar with ITK you may want to look at the Fill between slices. Unfortunately, this works only on image data so you would need to transform your data into a regular grid, with each polygon defining a segmentation. The filter will then generate a 3D volume from those. In the end, a Contour filter can be used to generate the output polydata that you want, something like a marching cube approach.

Another variant maybe to use vespa, which has a nice Alpha Wrapping algorithm coming from CGAL (Dual licensing, GPLv3 - Commercial). Using a large enough alpha, you would obtain a manifold polydata as output, enclosing all your slices in a 3D surface.

As for a VTK only solution, I am sorry but VTK is not really meant for meshing. Maybe using a linear extrusion to fill a part of the space between lines could help you get a better mesh, while still having a radius small enough to keep some concavities. In the end, there is no grantee the result will be manifold unfortunately.

1 Like

Hi, another option may be by using the PCL plugin that we have introduced in LidarView ( which, in short, is a ParaView based app meant to process point cloud data and connect to live 3D sensors).

We wrapped the PCL algorithms of Convex Hull, Concave Hull, Poisson and Marching Cubes to create meshes out of point clouds.

As of now, we haven’t isolated this PCL plugin from the LidarView build, so it has be to be built within LidarView ( but the resulting plugin shall be loadable in ParaView). Build guidelines may be found here