Equidistant Surface Points of STL Model

Hi everyone,

I am struggeling with a problem I thought would be not too difficult. I have a .STL model with its points and connectivities and want to get the surface as a equidistant pointcloud. The problem im having is that the points are only the edges of the connectivities and sometimes span larger areas than other times. (A square is for example only defined by its corner points while a dome is discretized into many areas)

I tried to generate a meshgrid of the geometry but this way not only surface points are generated but also points within a volume. I can also export the model with SolidWorks into another format but haven't had luck with other formats either.

Maybe somebody has a idea on ow to tackle this problem i'm unfortunately struggeling with.

Thanks in advance!

Lukas

Answer 1

[Maneet Kaur Bagga]

Hi Lukas,

• As per my understanding to generate an equidistant point cloud representing the surface of the .STL model, surface reconstruction can be used. This process aims to reconstruct the surface from a given set of points and their connectivities. Poisson reconstruction is one of the algorithm which can be used for surface reconstruction. It takes the point cloud and the connectivity information as the input and generate a reconstructed surface as output.
• After performing surface reconstruction, refine the point cloud by downsampling it to obtain an equidistant point cloud. It can be done by using voxel grid filtering.

Please refer to the following MATLAB documentation for implementation of the above methods:

Poisson Reconstruction

https://www.mathworks.com/help/lidar/ref/pc2surfacemesh.html

Voxel Grid Filtering

https://www.mathworks.com/help/nav/ref/statespacese2.sampleuniform.html

Hope this helps!

Regards,

Maneet Bagga