Why do our FR5 trim-plate scratch labels move with the light?

LuisBrown0956 · 7 Jul 2026, 23:09 UTC

Reply to discussion
LU
LuisBrown0956
Our proposed FR5 camera station presents brushed trim plates with small pose differences. Bright streaks shift, but their captions still call them scratches. I need a physical reference set and controlled image comparison before commissioning, including what happens when the image cannot settle the surface decision.

13 replies

LI
LinAllen0270
Replying to LuisBrown0956

Can each questioned caption be traced to the original capture and an identified retained trim plate? A screenshot can change the apparent feature before the camera setup even enters the argument.

10 points
LU
LuisBrown0956
Replying to LinAllen0270

Yes, the full captures and retained plates are linked. We have kept the captions as disputed. Quality has not judged the surfaces yet, so the links establish identity, not the answer key.

6 points
MA
MayaBennett0736
Replying to LuisBrown0956

What appearance method will quality use for the retained plates, so their viewing and cleaning conditions are consistent with the actual product requirement?

0 points
IS
IsaacAllen0346
Replying to LuisBrown0956

And preserve the accepted bright examples. A reference set made only of damaged plates can make the camera discussion forget how much normal finish varies.

9 points
LU
LuisBrown0956
Replying to MayaBennett0736

Quality used the agreed appearance method and accepted two broad streaked regions, but rejected a fine score on a third plate. The score has its own marked location; the bright band on that same plate is not the rejected feature.

15 points
IS
IsabelAli0213
Replying to LuisBrown0956

So keep the region decision with the plate identity, because calling the whole third image reject could teach somebody to use its harmless band as the reason it failed, even after quality has clarified the metal

5 points
LI
LinAllen0270
Replying to LuisBrown0956

Are those regions visible across the actual allowed presentation variation, or only in the images quality happened to annotate?

2 points
LU
LuisBrown0956
Replying to LinAllen0270

Only some existing views show the score clearly. Imaging and tooling are comparing its visibility at recorded seating conditions, alongside the accepted bright regions, before any threshold is selected.

21 points
MA
MayaBennett0736
Replying to LuisBrown0956

Does that comparison distinguish measured changes in seating from image movement alone? I want to know whether the holder is being assessed, not simply blamed because the stripe travels.

10 points
LU
LuisBrown0956
Replying to MayaBennett0736

Tooling measured the intended contacts and repeat seating; those are consistent within the reviewed presentation range. Imaging found the score loses contrast at one allowed angle under the current light, so the next candidate concerns illumination rather than a proven holder defect.

5 points
IS
IsaacAllen0346
Replying to LuisBrown0956

Useful distinction. The holder can be doing what was specified while the camera arrangement still cannot do the required inspection.

16 points
LU
LuisBrown0956
Replying to LuisBrown0956

The second illumination arrangement keeps the score visible in the small retained set at those angles and leaves the accepted bands distinguishable. We have not tested the full finish range yet; uncertain images still go to the defined quality review rather than becoming automatic passes.

19 points
IS
IsabelAli0213
Replying to LuisBrown0956

Keep the limited sample range in the result, Luis, and put the reviewed physical regions in the teaching material now; the old captions need not keep circulating while the wider imaging assessment is still unfinished

10 points

Add to the discussion

Welcome to Application Robot

Everyone can read the forum. Sign in or create an account to start a discussion, reply, or upload photos.

Forgot your password?

By creating an account, you agree to our Terms and Conditions and community guidelines. Read our Privacy Policy for how your information is handled.