Fast spherical polygon generation?

Posted 3 years ago
3665 Views
|
4 Replies
|
10 Total Likes
|

Cross posted on mathematica.stackexchange - https://mathematica.stackexchange.com/q/144167/5478

Question

Given a list of points on a sphere and the sphere/radius I'd like to plot a spherical polygon with vertices in those points.

And this needs to be fast, fast enough for the user to not "feel" generation time.

One should be able to style them too. Most importantly the surface but an edge style would be nice aswell.

What have I tried?

This is very closely related topic but answers there are not fast enough for my needs.

Is great but "only good for making spherical quadrilaterals, or isosceles spherical triangles".

• ClipPlanes in V11+ can be used as a directive which is very effective:
    RandomSeed[3];
pts = Normalize /@ RandomReal[{-1, 1}, {3, 3}]

Graphics3D[
{AbsolutePointSize@12, Point@pts,
Red, Sphere[{0, 0, 0}, .999], Blue,
Style[Sphere[], ClipPlanes -> {
InfinitePlane[{#, #2, {0, 0, 0}}],
InfinitePlane[{{0, 0, 0}, #3, #}],
InfinitePlane[{{0, 0, 0}, #2, #3}]
},
ClipPlanesStyle -> Directive[Opacity@.2, Red]]
}
] & @@ pts


But I'd need to write some code to determine in what order should those points be put in InfinitePlanes in order to clip from the right side (ClipPlane orientation). I didn't do this because I was too lazy and because:

> The number of clipping planes that can be implemented with ClipPlanes is limited by available graphics hardware.

So it won't be general enough. Though if you want to make this method automatic I will gladly upvote it.

Motivation

I think it will be useful in many applications.

I don't have time for this but I thought it would be a nice feature to have to improve code I was playing with lately, mostly based on another J.M.'s answer - Voronoi grid on a sphere

arc[center_?VectorQ, {start_?VectorQ, end_?VectorQ}] :=  Module[{ang, co, r}, ang = VectorAngle[start - center, end - center];
co = Cos[ang/2]; r = EuclideanDistance[center, start];
{{start, center + r/co Normalize[(start + end)/2 - center], end}, co}
]

points = {2 \[Pi] #1, ArcCos[2 #2 - 1]} & @@@ RandomReal[1, {10, 2}];

sp = Append[Sin[#2] Through[{Cos, Sin}[#1]], Cos[#2]] & @@@ points;

proc[] := (
ch = ConvexHullMesh[sp];
verts = MeshCoordinates[ch]; polys = First /@ MeshCells[ch, 2];

voro = Normalize[ Cross[verts[[#2]] - verts[[#1]],  verts[[#3]] - verts[[#1]]]] & @@@ polys;

edges =  arc[{0, 0, 0}, voro[[##]]] & /@      Select[Subsets[Range[Length[polys]], {2}],       Length[Intersection @@ polys[[#]]] >= 2 &];

);

proc[];

DynamicModule[{run = True},  Graphics3D[{ {Opacity[.75],
DynamicWrapper[EventHandler[Sphere[],
"MouseMoved" :>  Module[{pos = MousePosition["Graphics3DBoxIntercepts", True],
pt},  If[
Not@TrueQ@pos , pt = RegionIntersection[Sphere[], Line@pos];
If[pt =!= EmptyRegion[3],  sp[[-1]] = First@Nearest[pt[[1]], pos[[1]]]; proc[]]  ]]]
, TrackedSymbols :> {run}
]
}
, {AbsoluteThickness[2],
Dynamic[BSplineCurve[#, SplineDegree -> 2,
SplineKnots -> {0, 0, 0, 1, 1, 1},
SplineWeights -> {1, #2, 1}] & @@@ edges]}
, {Red, Sphere[Most@sp, .02], Dynamic@Sphere[Last@sp, .02]}
}
, PlotRange -> 1.1    , SphericalRegion -> True   , ImageSize -> 500]
]


To look more like https://www.jasondavies.com/maps/voronoi/

4 Replies
Sort By:
Posted 3 years ago
 I thought I had a simple solution, but I don't.
Posted 3 years ago
 I guess this comes originally from Demonstrations Project Voronoi Diagram on a Sphere by Maxim Rytin. I liked that you modernized it and interactivity works. It is broken in Maxim demonstration and I hope now they can fix it. Did you figure out what exactly the buy there was?But you also might be interested in another demonstration Voronoi Diagrams on Three-Dimensional Surfaces, which generalizes beyond the sphere and is by @Erik Mahieu who is a member here.
 I have not seen those demonstrations, thanks. The first one is probably broken as something in tetgenWolfram library has changed, I'd have to investigate. My approach is different, based on top level code provided by J.M. in the linked topic. The difference is it only generates points and edges that is why my example has no faces.About the second one, what it shows is an intersection of 3D Voronoi cells and given surface. I don't think it is the same as a Voronoi mesh on given surface. Maybe on a sphere. Or maybe it is, I'm not sure but I doubt it, would appreciate confirmation.
 Hi Sam,The implementation I gave in SE is essentially the same algorithm as Maxim's, where the relationship between the convex hull and the spherical Voronoi diagram was exploited. Maxim's old code used some TetGen functionality, which has now been supplanted by ConvexHullMesh, which I use in the code I wrote and Kuba expanded on.The code in the other demonstration you linked to does not seem to be using a geodesic-based distance, so I'm reluctant to consider it a true Voronoi partition.