how to calculate voxel image volume ?

Hi,

I am nex in matlab coding.

I am using voxel image function to drow voxels on a point cloud.

the code is runing and it gives a good plot. But the problem is that i do not know how to estimate the volume of my voxelized object ?

anyone can help ?

I already use volume function or [C,V] like alphashape or convexhull, but it doesn't work.

Thank's

[Walter Roberson]

is it https://www.mathworks.com/help/vision/ref/pcmapndt.html that you are using?

[Moussa DIEDHIOU]

Hi,

i am using voxel_image function : Voxel Image - File Exchange - MATLAB Central (mathworks.com)

the code is :

ptclouds = load('exemple.xyz');

Pts = ptclouds(:,1:3);

vs = [0.5, 0.5, 0.5]; % voxel size

vxl = voxel_image(Pts, vs);

Thank's

Answer 1

[Image Analyst]

Try regionprops3 or just sum the binarized image to count the voxels.

[Moussa DIEDHIOU]

can i have your adress to send you a short data to execute the code ?

I try regionprops3 and i can not have the same volume as convexhull or the terrain volume value. It suppose to be closer in my case because the cube fit well to the point clouds.

Excuse me because i am really new in matlab (only a few days) but i really love how it works.

Thank you

[Walter Roberson]

Image Analyst is not available through email.

[Moussa DIEDHIOU]

yes. Thank you for your help

Answer 2

[Walter Roberson]

discretize() each coordinate axis separately, getting an index number for each point. Use accumarray() with the combined indices, and count 1. Afterwards the volume is nnz() of the count times the volume of each voxel.

[Moussa DIEDHIOU]

i am executing thoose functions but it seems like i am facin some problems .

so, can i send the data to see how you do it ? .

Thank you

[Walter Roberson]

N = 500;
data = randn(N,3) .* [20 10 15];

ptc = pointCloud(data);

%pcshow does not work in MATLAB Answers

xyz = ptc.Location;

scatter3(xyz(:,1), xyz(:,2), xyz(:,3));  %instead of pcshow for Answers


minxyz = floor(min(xyz,[],1));
maxxyz = ceil(max(xyz,[],1));
dx = .5; dy = .25; dz = .5;

xidx = 1 + floor((xyz(:,1) - minxyz(1))/dx);
yidx = 1 + floor((xyz(:,2) - minxyz(2))/dy);
zidx = 1 + floor((xyz(:,3) - minxyz(3))/dz);
counts = accumarray([xidx, yidx, zidx], 1);

volume = nnz(counts) .* (dx .* dy .* dz)

volume = 31.2500

volshow(counts, 'BackgroundColor', [0.8 0.8 0.8]);

[Moussa DIEDHIOU]

It works ! Thank you so so much for your help Roberson.