Grain boundary smoothing

Stephane Dumoulin

unread,

Nov 12, 2019, 7:38:37 AM11/12/19

to dream3d-users

Hi,

I am trying to mesh on a microstructure. I used the example pipelines provided in the SimulationIO plugin. However, before meshing I use a smoothing filter (typically Laplacian). This gives fine results in the bulk of the volume. However, the grain boundaries at the surface of the volume keep their stair-like shape.

I have tried to play with the different options in the Laplacian filter but it tends to round off the corners of the volume as well as at GBs on the surface.

How can we smooth the grain boundaries at the surface without rounding them off?

By the way I see in the doc that there are other filters (e.g. "MultiMaterial Marching Cubes" or "Moving Finite Element smoothing") that seem interesting for the above purpose but they are not available in DREAM3D. I see those are in the repository while I see in another thread that e.g. M3C is not supported any more. Are you planning to fix them and include them back at some point?

Regards

Stephane

Michael Jackson

unread,

Nov 12, 2019, 3:00:58 PM11/12/19

to Stephane Dumoulin, dream3d-users

We do not currently have any plans to update the either of the filters as the original authors have moved on and the required expertise is no longer within the company. There is also another project that is close to being integrated that performs smoothinghttps://github.com/siddharth-maddali/HierarchicalSmooth that we are trying to integrate but we have dropped the ball on getting that one integrated.

The side effect of Laplacian smoothing is the rounding/shrinking of the volume. This is very hard to guard against in practice. We have not looked at the smoothing and meshing codes in a very long time. If we can secure funding in the future this would be a top priority area. In the mean time there are products just as SIMMETRIX that have added direct support for DREAM.3D data files into their products. SIMMETRIX is a commercial product.

--

Mike Jackson

--
You received this message because you are subscribed to the Google Groups "dream3d-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dream3d-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dream3d-users/0bc4fa99-8906-4f79-b2bd-6ba80501548a%40googlegroups.com.

Stephane Dumoulin

unread,

Nov 13, 2019, 3:29:27 AM11/13/19

to Michael Jackson, dream3d-users

Hi,

The hierarchical smoothing seems very interesting. We will probably look at it in more details and test it. Thanks for pointing to that one.

Regards

Stephane

Siqi Li

unread,

Nov 5, 2021, 7:07:51 PM11/5/21

to dream3d-users

Hello Stephe,

I was wondering if you have successfully solved this issue? Any help would be appreciated.

Thanks!

Best Regards,

Siqi

stephane...@sintef.no

unread,

Nov 15, 2021, 8:06:40 AM11/15/21

to dream3d-users

Hello Siqi,

We solve this by generating a seed map file that we used as input to another software (Neper) to generate a mesh.

Best regards

Stephane

Niklas

unread,

Nov 23, 2023, 12:36:43 PM11/23/23

to dream3d-users

Hello Stephane,

I'm currently working on a similar challenge, but in my case I want to generate a 3D-mesh by interpolating the grains in a coarse-grained sheet with a 2-sided EBSD scan. I'm now trying to evaluate whether its worth it to look into the approach you took of generating seed maps for use in NEPER.

Can you elaborate a little bit more on the approach and maybe share a picture of the results you got or a publication? Do you think it could be feasible in my case?