Figure 4. Workflow example. (A) Slice selector; (B) Cartesian cropping;
(C) cylindrical cropping; (D) filtered image; (E) and binary form.
With the aim of selecting the optimum threshold value for the glass-bead
sample, various calculations were performed on the 3D rendered image,
and information regarding porosity, number of clusters, and size of
clusters was derived. These calculations were carried out for each of
the considered thresholds and are plotted in Figure 5. For this sample,
a valley in the number of clusters vs threshold plot is observed
between threshold values 0.40 and 0.45. The porosity of the glass beads
calculated with Mathematica was found to be 0.378, at threshold value
0.45 located in the valley. On the other hand, Nakashima et al .
determined a gravimetric porosity of 0.370 for a glass-bead sample
[16]. These results indicate that a threshold value of 0.45 should
be optimum for the glass-bead sample. A glass-bead sample is known to
have only one pore system, and therefore should consist of a low number
of clusters in the PCA (the valley). The results of this experiment
indicate that the glass-bead sample can be successfully used as a model
to test the Mathematica code.