China Beverage MachineryLanguageEnglish
Vision Inspection False Rejects: Lighting, Presentation and Trials
Review beverage vision-inspection false rejects with sample truth, images, rule versions and separate missed-defect and rejection checks.
What to establish first
Trace false-reject reports by defect category, material lot and machine state. Review the image and reference criteria before changing a threshold; a change that improves one format can weaken inspection of another.
Three checks before a change
- Retain representative good, suspect and challenge samples. Record images, selected recipe, lighting conditions and the declared acceptance class.
- Compare position, rotation, reflections and material changes with the approved trial. Inspect optical cleanliness and setup through the authorised procedure.
- Re-run the agreed challenge set after a documented change. Confirm the reject action and tracking as well as the image classification.
Interpret the observed pattern
| Observed pattern | Next review |
|---|---|
| Rejects follow new artwork or format | Compare the approved artwork and inspection recipe with truth-labelled samples. A cosmetic change can require review without making every rejection false. |
| Rejects follow lighting or positioning changes | Reconcile maintenance, image conditions and actual package pose. Use a controlled comparison and retain images instead of editing rules from the aggregate rate. |
| Rejects fall after a rule change | Repeat the required good and defective sample set and verify physical rejection. A lower rate is not improvement if required defects are now missed. |
Evidence to record
Use these fields as a worksheet. Keep individual results and comparison samples; an overall average alone may hide the affected condition.
| Record item | Required evidence |
|---|---|
| Recipe and format revision | Identify the exact format, lot and document revision; attach a representative sample or record. |
| Good/suspect/challenge samples | State the location, method, instrument and operating condition so a comparison can be repeated. |
| Image and presentation conditions | Link the observation to an individual sample and a common timestamp; retain both normal and affected cases. |
| Detection and rejection verification | Record the approved assessment, action, verification result and responsible person; keep unresolved items visible. |
Controlled trial and acceptance
Agree the observation, affected boundary and acceptance method before starting a trial. Identify the product-package combination, approved settings and responsible equipment or quality owner. A test must have a clear question and a controlled disposition for trial material.
Compare the existing and proposed condition with identified samples using the same measurement basis. Change only what the responsible owner has authorised. Include the relevant steady-running, stop or restart case; one passing sample does not establish repeatable operation.
Keep results, samples and configuration revisions together. Record what passed, what remains open and the exact approved scope. If a correction moves the defect to another process step or compromises another acceptance requirement, retain that result and continue the assessment.
Common question
Can one vision recipe cover all label colours and bottle shapes?
Coverage needs evidence from the intended formats and materials. Similar appearance does not remove the need to qualify the relevant lighting, views and defect criteria.
Technical reading
The linked manufacturer material provides engineering context. This worksheet is an independently written diagnostic framework; the installed equipment’s approved instructions and the site’s product acceptance determine the actual procedure and limits.
China Beverage Machinery is operated by Allot Tech (Suzhou) Co., Ltd. for machinery sourcing, configuration planning and export project coordination.