Creating 3D surface from cross sectional curves

Hi guys!
I’m pretty new to Rhinoceros, and is running into the problem of using Loft command.
The objective of my project is to use 2D cross sectional curves to reconstruct the 3D surface model. The above problem arose because I have two curves in same elevation.
When I use Loft, Rhino creates surfaces that fills inside the gaps between the curves within same elevation. I also have this very ugly crease at the bottom which I don’t know how to resolve.
Please help me troubleshoot these problems. In advance, I appreciate any help and tips on how to create such model! Thanks!

Here is the link to the screenshots I took.

Hi Jasmine - that is a tough one to handle with a single Loft surface - NURBS surfaces do not lend themselves easily do construction of branching shapes - the four-sided structure that they have starts to get in the way. What you’d do is make the parts (three - the ‘base’ and the ‘branches’) and then, the hard part, figuring out how to bridge nicely across. If you post a 3dm file with the curves, we can probably get you pointed in the right direction.


Thank you for the input Pascal! I will definitely try your suggestion!
Here is the 3dm file that I uploaded to because it won’t let me upload files here somehow D:!k9JRjbiB!BZ1cga2IZK15moOwnPN_EVMbY_GACfpoXT_inqjxCJk

This might be a weird question. But is there any good software out there that handles this kind of situation well? Because for the next few weeks I will only be using CAD software for this purpose. Thank you guys for helping guy :smiley:

Yeah… this will not be easy with just this input - what is it, exactly? Is there any way to get more of an idea of what you want it to look like?


Hi Pascal,

So the curves you see are actually traces of the MRI spine images.
My goal here is to take individual MRI image, manually trace out the spine cross section, and try to reconstruct the actual spine by stacking the curves on top of each other and create the surface of the spine.
Since the surface will be quite complicated as the bone surface is not uniform, I’m really hoping that Rhino would not make the surface too “curvy” and “spiky” with its Loft algorithm.
I’m testing out Rhino because I read a paper that used this software to achieve similar purpose. Their result is great and their surface is smooth, so I’m guessing I must be missing something here!

A single spinal vertebrae looks like this, so this is something I try to reconstruct.


Hi Jasmine- the problem you’re likely to encounter is that surfacing wise, defining a shape directly using a series of parallel sections is seldom a good way to proceed. In this case, the additional problem is that your sections are simply too sparse to really be useful in understanding a complex branchy shape like the one in your image. I believe you’ll want to use a SubDivision modeler of some kind or conceivably TSplines (plug-in for Rhino) to make this sort of shape at all efficiently.


1 Like

Hmm, interesting. I’ll take a look into that plug-in and hopefully it helps. ty man!