Implement a three-dimensional Convex Hull algorithm to process voxels at the surface of voxel models from Goxel imports

[andrewrgarcia]

Some 3D models may have internal voxels which are not seen and are thus a waste of memory, computation-wise.

A 3D Convex Hull algorithm will process the external voxels which are visible alone.

This will reduce the time complexities of rendering and dynamic manipulation of 3D objects with "n" lengths from O(n^3) to O(n^2) for cases where internal voxels are present.

While a Jarvis algorithm wraps all points by considering the most counterclockwise trios:

A 3-D Convex Hull algorithm will generate a hull from an initial triangle plane of 3 voxels and pivot around the external points in 3-D space until the hull closes:

Some useful references: https://www.cs.jhu.edu/~misha/Spring16/09.pdf

[andrewrgarcia]

It is currently being developed as MeshWrap in voxelmap versions >= 2.1.0 see source-code

[andrewrgarcia]

Current issues with MeshWrap

• graphical bugs Will not do triangulation on smooth surfaces. This is intrinsic to the 3-D Convex Hull algorithm; must either set up a constraint on smooth sites or make smooth sections rougher with an image manipulation technique
• automation MeshWrap is far from automated. Must manually define some constraints such as for 3-D spacing [ for wrapping ] rel_depth and 3d resizing voxelmap.resize3d to make adequate Triangulated objects.
• documentation It currently lacks documentation.

Update: MeshWrap has been supplanted with MarchingMesh function

[andrewrgarcia]

This issue has been resolved with the MarchingMesh() function; applying a marching cubes function from scikit-image.