Our accepted and rejected faceplates are identified, and quality has marked the actual rejected feature on the retained part. The FR10 presentation trial still gives us wildly different bright bands. People are calling every strong band another scratch.
Looking through the capture notes, exposure is automatic and nobody recorded the support position. I don't want the vision person tuning around a collection of unrelated pictures. Which small comparison should we rebuild first?
Repeat the same identified faceplate with the vision person recording the imaging settings and presentation. Can the support reproduce its position, or does it move between those captures as well?
Support position is repeatable once located. The draft comparison mixed located captures with handheld ones, and the filenames didn't distinguish them. Originals are still there.
One handheld accepted example and one located rejected example. So the threshold comparison had changed the presentation method as well as the part. I've taken that comparison out of the proposal.
David, retain the awkward accepted surface in the new comparison. Controlling the capture shouldn't become permission to choose only an easy accepted plate.
Ravi, the required edge is exposed in the located view. Imran, the same accepted plate stays in. Vision has repeated both under one recorded setup; its strong band remains, but the rejected feature is weak in that lighting.
Then what alternative lighting comparison is proposed? The corrected capture method has exposed a limitation worth addressing before the purchase decision.
They want to compare another lighting position with both plates and the same located presentation. I've added that outstanding comparison to the proposal. No new camera order and no threshold chosen.