Hey all! Have I missed the right way to convert high-poly mesh into a single NURBS polysurface? Packed NURBS object isn’t helpful, because I can’t merge the faces. Must be further editable, mesh and SubD don’t suit me.
Also tried GH to convert each mesh face to NURBS, but still don’t see how to merge them together.
Would you suggest any solutions, please?
There are entire programs dedicated to reverse engineering. I fear there is no easy way to do with 3 clicks… probably the easiest that I know at the moment would be xNurbs, althought most things can be done without just with Rhino. Depending on the complexity of the file, you wil need to subdivide the mesh into surfaces, that make sens for you and the geometry. If you post the file, the community can usually help better than help for general cases.
As Ben said, Reverse Engineering is an entire field un to itself. There is no general 3-click-solution. The way to get the best result is just to model it from scratch using the mesh as a reference.
What is the overall shape? What does the mesh look like? The best approach may be to use the mesh to model a new surface, which is what I usually do for boat hulls.
Having the mesh file for reference would be helpful as others have said.
Also relevant is how the file will be used … is it for manufacturing …molding, CNC, or other???
Thanks a lot for your replies! Here is the mesh, original is almost 25K polygons. It can be reduced, but every mm matters, so the geometry can’t be simplified significantly. It’s very important to keep this bottom geometry, so I didn’t like the vertical loft results either.
It will be a 3D printed object after I trim it, offset and blend another nurbs object on the bottom.
short answer, you won’t get a single surface… it will be a polysurface any way you slice it.
quadremesh>tosubd>tonurbs is the current workflow for reverse enginereing in Rhino, but the result will be a polysurface, not a single surface.
if you are going to print it, it will be stl anyway, so what you can do is keeping the object as mesh as reference, just rebuild the relevant surfaces and build on these. then boolean union. usually printers nowadays check the files and unify them anyway before slicing, so you probably don’t even need to unify the objects. it then is ready for export to stl…
it’s a workflow I often use (right now in fact) that avoids you rebuilding the entire thing.
Many thanks for your suggestions! 
this is very good advice.
