Jeff, it's really simple, actually. You use the color balance sliders in a photo editor to eliminate the colors you don't want.
As you know, most camera sensors use RGB filters to isolate the color spectrum at adjacent pixel locations (most use a "Bayer pattern", see Bayer Filter), which are then combined by the software to determine the color at each pixel in the final image.
If one were a total purist, you would capture in RAW and only use the pixels from a given color for each image. In practice, that's not necessary. The info in the camera's .JPG file is quite sufficient.
I use XNVIEW (see Xnview homepage), it is freeware and has several features which are very handy for this sort of work. And the author has had, from day one, an obsession with accuracy in his routines. Here's a screen shot of the green being isolated by killing the blue and red channels.
Referring to the images in post #30, you can tell that the blue focus is noncoincident with the red and green, because its Ronchi lines are spaced differently. That's what creates the colors in the original "white" image. In the out-focus images, the blue is closer to the focus point (line spacing is further apart), while in the in-focus images the blue is further from the focus point (the lines are closer, and more of them).
I doubt if we can use this technique to accurately "measure" the focus point difference. But it can certainly detect the differences.
Edited by precaud, 16 February 2020 - 12:47 PM.