OpenMV interface detection

I’m wondering if OpenMV library has existing CV functions that would be capable of detecting a liquid-liquid interface (say oil in water) and outputting the length of each interface (from a defined fixed length in the frame) say for example from the bottom/top of a tube, then use some sort of confidence interval or some other means for determining approximate time for interface to develop.
If anyone has any input on how to go about implementing this, I’d gladly accept as well, but mainly just wondering if something like this would be possible without custom functions or if I’d be better off looking at using a generic cam with openCV or other software.

You should checkout Edge Impulse and train a CNN to do this. Also, your question is vague. With CV you need to be explicit. Computer Vision is not magic. You should have some type of method by which you plan to accomplish your goal.

Appreciate the input.
This is my first foray into CV so vagueness is probably more of a result of lack of knowledge rather than intention, so using not CV lingo the gist of my goal would be.
I think I’ve found something that will work… image.find_lines()

Camera would be setup to point at side elevation of a sightglass of a predefined fixed ‘reference’ length.

Fluid would enter sightglass and be allowed to settle/separate. The fluid interface should develop a ‘line’ that should be detectable. A repeatable settling (line detection time) is more important than getting an absolute ‘trained’ detection at a specific point in the phase separation. Alternatively, you could look for two rectangles instead of line, color interface, etc…

-Define cordinate boundary
-Trigger Output (fluid fill)
-Start Timer/Loop
***Look for horizontal line in boundary (img.find_lines…???)
-if line found, stop timer. end loop

Settling Time = Stop Time - Start Time
Heavy Phase = (Y Line - Y bottom bound)/(Y top bound - Y bottom bound) * boundary length
Light Phase = (Y top bound = Y line)/(Y top bound - Y bottom bound) * boundary length