voxelize | Voxelizes a triangulated mesh into an ndarray | Service Mesh library

I'm doing a small project which requires to resolve a bug in matplotlib in order to fix zorders of some ax.patches and ax.collections. More exactly, ax.patches are symbols rotatable in space and ax.collections are sides of ax.voxels (so text must be placed on them). I know so far, that a bug is hidden in draw method of mpl_toolkits.mplot3d.Axes3D: zorder are recalculated each time I move my diagram in an undesired way. So I decided to change definition of draw method in these lines:

I have a point cloud data(x,y,z), I want to voxelize this point cloud and get all the voxels. PCL voxelgrid implementation voxelizes the point cloud and return the centroid of each voxel. Exact implementation of voxel grid is in the url 'http://docs.pointclouds.org/1.8.1/voxel__grid_8hpp_source.html#l00214'. Any other references for voxelizing the point cloud is also appreciable.

What is the performance of transforming between geometry primitives of polygons and voxels by VoxelizePolygons?

Is there any benchmark tests on it? I'm trying to estimate if it is a critical overhead for an application handling polygon meshes of around several hundred mega bytes in average size.

So I am building a voxel-based physics simulator in which voxels can be destroyed. Every voxelized object has a kind of "center voxel" to describe it, ill call it "peace A".
All voxels are simulated as one until they are separated from "peace A" or another voxel attached to it, then they(and/or other voxels connected to that voxel) get put into a new object with its own simulated physics. And here lies my problem, How Do I check if the voxel is still attached? in the most efficient way possible?

pathfinding sounds like it would slow down allot when I scale up but happy to try and figure that out if its the best option. and updating all voxels at once when one voxel is destroyed doesn't sound too good either. Any of you geniuses have any ideas?

here are some drawings for clarity(ps please excuse my bad mouse-handwriting)

first image, with voxels still connected:

second image, when the voxels separate:

I am working on a octree traversal algorithm. The current implementation uses a std::queue for such purpose, working flawlessly. However, I would like to use for such traversal a std::stack, as a depth first search will give better performance, avoiding testing non needed nodes.

However, when changing from one structure to another, I start getting segmentation faults on the push() function. Here is the stack report from gdb:

This seems like a fairly simple problem, and I haven't been able to find a canned solution. Essentially, this would be the 3d equivalent of skimage.measure profile_line

Consider a matrix A, with dimensions (i,j,k). Each matrix element is a number. In real terms, consider this a voxelized 3 dimensional temperature distribution.

I would like an efficient method to extract a line profile (a ray trace line) through this data from point (i_1,j_1,k_1) to (i_2,j_2,k_2). Or, similarly, defining an initial point (i_1,j_1,k_1) and a line trajectory using theta, phi and radius in polar coordinates.

I recognize an accurate result would consider boundary crossing and partial voxel volumes along the path, but I would be satisfied with a crude approximation which samples nearest voxel values with a regular step size (say, 0.1*voxel dimension) along the ray trace line.

Help is much appreciated. I am happy to describe further as needed,

I call the kernel function like this

Im currently in the process of writing a Voxel Cone Tracing Rendering Engine with C++ and OpenGL. Everything is going rather fine, except that I'm getting rather strange results for wider cone angles.

Right now, for the purposes of testing, all I am doing is shoot out one singular cone perpendicularly to the fragment normal. I am only calculating 'indirect light'. For reference, here is the rather simple Fragment Shader I'm using:

I would like to turn a 3D point cloud into a simple 2D occupancy grid. In my current implementation I first apply the voxel grid to a point cloud and then I manually iterate through all points in the voxelized cloud to fill in my grid. Is there a way to make it so that while I am voxelizing the grid, I am generating my 2D occupancy grid?