isSelfIntersecting function gives apparently incorrect result

5 views (last 30 days)
I'm trying to use the isSelfIntersecting function from the Lidar Toolbox to check whether or not a surfaceMesh object is self intersecting. In most cases it works fine, but in certain situations it returns true when the surface is not self intersecting.
The problem seems to arise in situations where all vertices in the mesh are contained, or nearly contained, by one plane (as in the example below). I wondered if it might be an issue with leaving the face normals undefined, but I get the same issue even if I define them when calling surfaceMesh (not included in example for readability).
Is this an issue with the way I'm defining the mesh and using the function? Or is this a known issue with isSelfIntersecting?
Many thanks for any help!
% define the vertices for the mesh
vertices = [0 0.0785926524819716 0.0150000000000000; ...
0.00223394141607314 0.0787352254127876 0.0140425952680485; ...
0.00428764473102951 0.0788687035475108 0.0131624353567936; ...
0 0.0822253462122803 0.0150000000000000; ...
0.00224235572430482 0.0822979726488426 0.0140389888997977; ...
0.00430286261362254 0.0823659662099516 0.0131559129739490; ...
0 0.0860000000000000 0.0150000000000000; ...
0.00225109241982507 0.0860000000000000 0.0140352443565181; ...
0.00431866355685131 0.0860000000000000 0.0131491406913786];
%define the faces for the mesh
faces = [1 2 5;
1 4 5;
2 3 6;
2 5 6;
4 5 8;
4 7 8;
5 6 9;
5 8 9];
%create the surface mesh object
mesh = surfaceMesh(vertices,faces);
%plot the mesh and verify visually that it does not self intersect
%(it should be an inclined plane sloping downward along the x-axis)
patch('vertices',vertices,'faces',faces,'EdgeColor',[1 .5 0],'FaceColor',[.5 1 .5]);
view(3)
%call isSelfintersecting, returns true despite mesh appearing to be
%non-self intersecting
isSelfIntersecting(mesh)

Accepted Answer

Deep
Deep on 27 Jun 2023
Hi Ed,
No, you have defined the mesh correctly and are using isSelfIntersection just fine. This is likely due to floating point errors. On inputting your vertices data, it appears slightly differently in MATLAB workspace variables due to the precision limitations of double.
If you look closely, you can see some differences from your original input. For example, in the 2nd column 1st row, there's a diff of 4e-16. Unfortunately, this seems to be a limitation with floating point computation. Hope this helps!
  3 Comments
Deep
Deep on 27 Jun 2023
I am not aware why that would happen. This might require more information about your system setup and these scripts. Perhaps the output was copied from a computer with a different hardware platform? You may create a new question with more information about these scripts so it can be identified why this is happening.
Ed Wheatcroft
Ed Wheatcroft on 28 Jun 2023
Edited: Ed Wheatcroft on 28 Jun 2023
Ok, no worries :) I actually found another configuration which gives an apparently erroneous result while I was playing around with this. This time, the mesh is clearly self intersecting, but isSelfIntersecting returns false.
In this new example the entries in vertices should be exactly representable using double (or indeed single) precision, so I wonder if this is even a separate issue.
Either way, it seems like isSelfIntersecting isn't totally reliable.
vertices = [0 0 0;
1 0 0;
0 1 0;
0 0 1;
0.5 0.5 1;
0.5 0.5 -1];
faces = [1 2 3;
1 2 4;
1 4 5;
1 5 6];
mesh = surfaceMesh(vertices,faces);
patch('vertices',vertices,'faces',faces,'EdgeColor',[1 .5 0],'FaceColor',[.5 1 .5]);
view(3)
isSelfIntersecting(mesh)

Sign in to comment.

More Answers (0)

Products


Release

R2023a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!