简体中文界面。帖子、指南及政策正文保留原文;未对原文作自动翻译。

Why do our FR10 faceplate examples still call moving streaks scratches?

PriyaAdams0133 · 2026年6月21日 18:07 UTC

回复讨论
PR
PriyaAdams0133
On another faceplate review, quality found the actual scratch away from the moving stripe. This FR10 camera plan still has streak-based labels and unrecorded presentation changes. Before commissioning, how do I separate usable physical references from images that only show uncertain appearance?

3 条回复

PA
PavelAli0179

Keep sample identity, physical region and quality judgement linked. Retain the uncertain images separately with their presentation conditions where known. A result from your other review cannot be assigned to these examples merely because the streak looks similar.

3
PR
PriyaAdams0133

Agreed. I've removed the confirmed-scratch status from the unverified examples without calling them accepted. Two can be matched to retained faceplates; the others still lack a physical identity. Threshold tuning is on hold.

6
PA
PavelAli0179

Have quality review those two identified faceplates against the applicable requirement while fixture and vision owners examine presentation repeatability. Keep unmatched images as observations only. They cannot become acceptance references through a folder rename.

5

参与讨论

欢迎来到 Application Robot

所有人都可以阅读论坛。登录或注册后即可发起讨论、回复或上传照片。

忘记密码?

注册账号即表示你同意我们的 使用条款 社区准则。请阅读我们的 隐私政策 ,了解我们如何处理你的信息。