I worked on a similar subject in 2D (https://doi.org/10.1016/j.matdes.2015.12.062). As far as I know, you may encounter two problems:
- First, you may get “parasite” branches due to local geometrical irregularities of the struts (diameter variation, protuberances, etc., especially if it’s additive manufacturing). These create branches that do not correspond to struts.
Analyze skeleton can identify them, though, or even prune them. Or you may want to “smooth” your volume (e.g. by mathematical morphology) before skeletonization.
- Second, it may happen that branches corresponding to connecting struts do not intersect at a single quadruple point (I guess your structure is diamond-like), but at two or three close points. If so, you will have to find a way to “gather” these junctions into a single one representing the geometrical center of a node.
And I think you should clip the top and bottom of your volume because the skeleton is ugly there (maybe because you have upper and lower plates).
But first, as iarganda said, use
analyse skeleton, and check what you get.