Moving streak in the bezel references

Start with that question. I've seen a perfectly real machining mark get confused with its much brighter reflection, then a new light made the reflection disappear and everyone declared victory. Keep the physical location visible in a wider image. Have quality identify an accepted part and a genuine reject, then reseat both in the intended holder. Don't keep moving the arrow to whichever bright thing looks persuasive that day.

The reject label came from my draft worksheet. Quality had not rejected that bezel. I have removed the label and asked them to select physical examples against the appearance requirement.

Will learners handle those examples, or only see photographs?

I'd let them compare the parts with the pictures if the lesson permits it. Not asking them to become inspectors, just making it obvious why a wandering highlight is a poor landmark.

They can handle the teaching samples. Quality has supplied an accepted bezel and one with a rejected scratch, with the judged areas identified. The vision engineer is comparing both through repeated seating, using the existing light first.

Did the real scratch stay visible after the lighting change? That is the result I'd want beside the improved image of the accepted part.

Yes. Repositioning the existing light reduced the wandering band without losing that scratch in the repeated seating comparison. I should say reduced, not removed: a weaker band remains on the accepted bezel. That is useful for the lesson.

We ran the revised exercise with the physical samples. One learner initially chose the brighter accepted bezel, then found the actual scratch by matching the part to the full image. The worksheet now asks for that match before a judgement. No replacement light ordered.