A scan boundary is not always the part surface
If a specimen extends beyond the scan, the reconstruction edge can be closer to a pore than any observed physical surface. The boundary setting changes the meaning of the measurement.
Input and calibration
A 30 × 30 × 30 volume contains material at [:, 5:25, 5:25], so the part crosses both Z boundaries. A 3 × 3 × 3 cavity lies at [2:5, 14:17, 14:17], in Z,Y,X order. The material intensity is 1000, air is 0 and spacing is 1 µm on all axes.
Method
Run pores_core.analyse twice with threshold 500. Keep the input and segmentation settings fixed; change only include_fov_cut from true to false.
Computed results
| Boundary setting | Surface distance (µm) | Nearest target |
|---|---|---|
| Include cut | 2 | Field-of-view cut |
| Exclude cut | 8 | Part surface |
Including the scan cut gives 2 µm and flags the nearest target as a field-of-view cut. Excluding the cut gives 8 µm to an observed part surface. Neither setting reconstructs the unseen specimen outside the scan.
What this demonstrates
This is a controlled example, not a reconstruction of missing material. Report the boundary setting with the result, and assess whether the observed scan contains the surfaces relevant to your question.
Reproduce the geometry
Construct the two-level volume from the array indices above, use the stated voxel spacing and threshold, and apply the same boundary setting in Pores to Surface Analysis. See the linked guide and manual for the application workflow.