In my scenario, I have a focus stacking issue. I'm trying to do a scan of a PCB which is very thin in one axis but thick the in others. Broad value Interval-based focus stacking simply doesn't work in this scenario as it results in very fuzzy images when taking images perpendicular to the thin axis. After doing some high level research into exactly what a diopter is think I may have a solution for this:
For each autofocused image, record the resulting diopter, theta, and phi.
Turn each diopter into an actual focal distance by simply inverting it (1/diopter value).
Use the focal distance as a "virtual" vector projected from the angle and position determined by theta and phi.
Create a mesh/boundary using the endpoints of those vectors.
Use the mesh/boundary to determine the near and far focal planes used for focus stacking.
In my scenario, I have a focus stacking issue. I'm trying to do a scan of a PCB which is very thin in one axis but thick the in others. Broad value Interval-based focus stacking simply doesn't work in this scenario as it results in very fuzzy images when taking images perpendicular to the thin axis. After doing some high level research into exactly what a diopter is think I may have a solution for this:
For each autofocused image, record the resulting diopter, theta, and phi.
Turn each diopter into an actual focal distance by simply inverting it (1/diopter value).
Use the focal distance as a "virtual" vector projected from the angle and position determined by theta and phi.
Create a mesh/boundary using the endpoints of those vectors.
Use the mesh/boundary to determine the near and far focal planes used for focus stacking.