Surface from Metaballs


first of all i know that this topic has been discussed a few times already but i couldn’t find a solution for my problem.
In our designing class we are creating cells for moss to grow on a facade. So we used a microscope picture to get a surface. From that we extracted single cells by just cutting the surface. We can basically work with that and just offset it as cells.
On that surface we applied the shortest walk and got the points from that curve by dividing it.

An with that points we got that form from the metaballs which look way nicer than the simple cells we get from just cutting the image sampler surface. But that is also the part where the problem starts.
We have the curves but we can’t get a surface out of them.

I already looked at:
but i couldn’t get the meshmachine to work (it worked with a simple box but not with the mesh i get)

didn’t work and the last time i tried rhino/grasshopper crashed

didn’t work either

helped a lot but the mesh from points doesn’t look good

also looked here
Panelling metaballs - Grasshopper (

and tried everything from here
Metaballs to mesh in year 2020 - Grasshopper - McNeel Forum

I guess one of the method would work if we just had some metaballs but bot als walls of a cell.

That’s the result with the ball pivot and the points i got from the method with 3 planes and the ball pivot (described by Laurent Delrieu)

Honestly i can’t remember what i tried here but it’s not what we want anyway cause it looks more like a blown up version of our original cells than the version we get with the metaballs.

Tried with the dendroballs. Yeah the size of the balls is to big but i also tried with smaller ones and it looks similar to the ones we got abouve. So also not what we are looking for. Here we used the same points as for the metaballs.
Also tried dendro with the curves but that doesn’t look good either.

Left: got the points by dividing the curves resulting the metaballs with 3 planes.
Right: points are from our original metaball.
(the curves were baked and divided by length in rhino)

Also with the points extracted from the curves from the metaballs with one plane.
Used aroun 300000 points for these two and the next one. Just with different SamplingDensityPlusNoise

Has anybody another idea what we could try or is it just a lost cause and we should leave it.
We would especially like to keep these areas. It doesn’t have to be the fastes way… waited for some tried also 10 min but if it works i don’t have a problem with that. We just need 2-3 different version of the cell so that would be ok.

If it helps my grasshopper file. there is almost everything in there we tried. (134.5 KB)

Sorry for the long post and thanks in advance.

You can cull out the duplicate Metaball lines, create boundary surfaces then extrude the surfaces down to the adjacent level. This will give you an architectural topographic style solid from your curves.

From here you can bring that set of bodies into Dendro to do your union and smoothing. Finally, offset your volume back out a bit to recover the lost area from your smoothing operation.

Not perfect and you may have to fuss with your surface division count and Dendro settings a bit but it should maintain the Metaball look and give you a solid mesh in the end.

1 Like

That’s perfect. Thank you such much!
You don’t know how much you helped :hugs: