Kangaroo surface planarization

Replies to This Discussion

Permalink Reply by mia on December 27, 2015 at 5:46am

hI Manu there is already a planarization component in kangaroo latest version.

Permalink Reply by MANU on December 27, 2015 at 10:21am

thnx for the reply kartik.... but  i m still confused... .. because i have used that tool which u have mentioned and nothing is happening.....

can you tweak d definition little bit...

Permalink Reply by Joseph Oster on December 27, 2015 at 1:19pm

Hello MANU.  Why Kangaroo?  I thought you had this voronoi surfaces stuff all worked out in the other thread?  Voronoi pattern on an object

Or is the key word here planar voronoi surfaces?  How can surface segments be planar unless they are triangles?

The first thing I noticed looking at your code in this thread is the severe performance penalty caused by flattening the curves input to 'SrfSplit' - 2.5 minutes!  Ouch.  Without flatten, 'SrfSplit' takes less than half a second.  So I started messing around, applying some lessons from our earlier adventures, copying the code you wrote (in the blue group) for facets - though I modified it slightly to eliminate one more 'Area' component to get a centroid point for 'Scale'; I used the point we already had from 'Srf CP' / 'EvalSrf' to scale each surface BEFORE the 'Move' with surface normals.  I also applied 'Cls' (test for closed curves) prior to 'SrfSplit' to ignore the edge segments.

I am very puzzled, though, to see that the facets don't sit well on the original surface but pass through it instead?  I tried the old code with this surface and got the same result - very strange because I don't see this problem with any of my other test surfaces in the old code...  It's not noticeable (the holes look fine) until you show the original surface and then the difference becomes apparent.  'Project', 'Pull' and 'BBX' all produce the same result.

Attachments:

• voronoi_2015Dec27a.gh, 34 KB

Permalink Reply by Joseph Oster on December 28, 2015 at 11:20am

Hello again MANU.  Much to my surprise, I came across an example of planar voronoi surfaces on a hemisphere (without Kangaroo), in this thread: morph on surface

I don't know if it's possible with your surface here but two things to note:

1. The points to create the voronoi curves are not random.  They are carefully culled from circular patterns that distribute them evenly.
2. The curves and points are not projected but mapped.  The surface used by 'Srf CP' in the other thread isn't just a circle/boundary surface, it is a 'NetSurf' created with "spokes" and concentric circles to get the 'uvP' points used from 'Srf CP' and mapped to the hemisphere by 'EvalSrf' for the polylines that make the planar boundary surfaces.

I'm not sure yet how to apply it on the tent-like surface you have here, or if it's even possible.  Notice how evenly space they are on the hemisphere: