The acceptable plate is still acceptable under quality's check. Its latest camera image has a bright band labelled scratch, and the band shifts when we reseat it. I am not the imaging specialist, but this seems a poor starting point for tuning anything. Who should settle the disagreement between the physical reference and the annotation?
Have quality and the imaging person look at that exact plate together. What does the arrow point to on the physical surface? Keep the unmarked image too; nobody needs to defend an arrow just because it was saved first.
Nothing corresponding to the band was found on the plate. Quality withdrew the scratch annotation. The imaging person wants a controlled presentation comparison before changing the light, which seems reasonable.
Did reseating explain the band change? I'm interested whether the holder itself lets the plate sit differently, rather than assuming the robot would reproduce the hand-held pictures.