I used a he default volume bounding algorithm of the Spc.Combine function to fit a boundary sphere around a dodecahedron.
It's naive algorithm then calculated the size of a cuboid fitting around the object, then fitted a sphere around that! Works great on actual cubes, but was pretty lousy for my dodecahedron.
In than case it seemed to construct a bounding sphere of radius = 3, rather than the best fit radius = sqrt(3).
Can this naive approach be easily improved..?
I used a he default volume bounding algorithm of the Spc.Combine function to fit a boundary sphere around a dodecahedron.
It's naive algorithm then calculated the size of a cuboid fitting around the object, then fitted a sphere around that! Works great on actual cubes, but was pretty lousy for my dodecahedron.
In than case it seemed to construct a bounding sphere of radius = 3, rather than the best fit radius = sqrt(3).
Can this naive approach be easily improved..?