Hur man automatiserar ISBM-flaskkvalitetsinspektion

What this article must prove

Automate bottle quality inspection by defining defects first, selecting cameras and sensors for the optical task, controlling bottle presentation, linking cavity traceability and proving false-reject/escape performance. A defensible baseline begins with Collect real examples of black specks, haze, gate marks, flash, short blow, deformation, scratches and contamination. The first verification method is Label by severity and customer specification. From there, the article follows only checks that can materially change the answer promised by the title. Where an exact operating value depends on the resin grade, bottle drawing, mold, or delivered machine, the approved project specification controls the final setting.

Define What the Vision System Must Decide

Defect library

Defect library. Collect real examples of black specks, haze, gate marks, flash, short blow, deformation, scratches and contamination. Label by severity and customer specification. For defect library, define the defect or measurement decision first and then choose sensors, lighting, or logic that can prove it. For this topic, the engineering log should connect defect library with the observed bottle condition and then test whether defect library supports the same diagnosis. Buying a camera before defining defects leads to impressive images but weak decisions.

Choose Inspection Points from Actual Defects

Inspection timing

Inspection timing. Choose whether to inspect immediately after molding, after cooling, after leak testing or before packing. Place each check where the defect is stable and bottle presentation is controlled. For inspection timing, define the defect or measurement decision first and then choose sensors, lighting, or logic that can prove it. Do not judge this factor from the HMI value alone; pair it with the actual bottle result, then continue with inspection timing only after the relationship is clear. A warm bottle may change shape after the vision decision.

Control Bottle Presentation and Lighting

Lighting

Lighting. Use backlight, bright-field, dark-field or polarized arrangements for transparent bottle defects. Test the real resin colors and surface gloss. For lighting, define the defect or measurement decision first and then choose sensors, lighting, or logic that can prove it. Keep lighting at its validated baseline while this item is tested so the bottle response can be attributed to one cause. One lighting setup rarely reveals both black specks and subtle surface scratches equally well.

Bottle orientation

Bottle orientation. Control rotation and position for oval bottles, logos, gates or asymmetric panels. Use mechanical guides or tracking features and verify repeatability at production speed. For bottle orientation, define the defect or measurement decision first and then choose sensors, lighting, or logic that can prove it. Use the smallest controlled change that can prove the effect of bottle orientation, then restore the baseline before a different adjustment such as bottle orientation is tried. Random orientation causes false rejects when the system compares the wrong region.

Working terms for this specific task

Defect library

Collect real examples of black specks, haze, gate marks, flash, short blow, deformation, scratches and contamination.

Inspection timing

Choose whether to inspect immediately after molding, after cooling, after leak testing or before packing.

Lighting

Use backlight, bright-field, dark-field or polarized arrangements for transparent bottle defects.

Bottle orientation

Control rotation and position for oval bottles, logos, gates or asymmetric panels.

Inspect Neck, Body, Base and Gate Separately

Neck inspection

Neck inspection. Check thread/snap finish, flash, ovality indicators and contamination using suitable optics. Combine vision with gauge or dimensional sensors when absolute dimensions are critical. For neck inspection, define the defect or measurement decision first and then choose sensors, lighting, or logic that can prove it. This factor belongs in the setup sheet because it directly changes the conditions under which neck inspection is evaluated. A visually clean neck can still be dimensionally out of tolerance.

Add Dimensional or Leak Tests Where Vision Is Not Enough

Base and gate

Base and gate. Inspect gate centering, black specks, cracks, base deformation and foot geometry. Use multiple camera angles if deep push-up features hide the gate. For base and gate, define the defect or measurement decision first and then choose sensors, lighting, or logic that can prove it. If the result differs by cavity, compare the local hardware related to base and gate before moving on to base and gate. A single top camera may miss a defect on a recessed base.

De multi-material ISBM platform overview also helps frame how machine architecture, materials, utilities, and bottle applications fit together at line level.

Link Defects to Cavity and Recipe

Body inspection

Body inspection. Detect color, haze, short blow, panel deformation and foreign particles within the practical limits of the optical system. Teach the system using real good/bad distributions rather than perfect CAD images only. For body inspection, define the defect or measurement decision first and then choose sensors, lighting, or logic that can prove it. This item is considered resolved only when the finding remains repeatable after thermal stabilization and the next check, body inspection, does not contradict it. Transparent walls and reflections can create false edges that confuse simple thresholding.

Cavity traceability

Cavity traceability. Associate each bottle with mold cavity where machine and handling sequence allow it. Trend defect rate by cavity. For cavity traceability, define the defect or measurement decision first and then choose sensors, lighting, or logic that can prove it. For repeatability, define who measures cavity traceability, where it is measured, and what bottle evidence is required before checking cavity traceability. Vision without cavity data identifies scrap but misses the opportunity to repair the source quickly.

Design Rejection and Verification Logic

Reject mechanism

Reject mechanism. Use a confirmed reject action and sensor to verify the bad bottle leaves the good stream. Design for the maximum line speed and bottle variation. For reject mechanism, define the defect or measurement decision first and then choose sensors, lighting, or logic that can prove it. Record the bottle response beside the setting or measurement for reject mechanism; that record becomes the starting condition when reject mechanism is reviewed. A correct inspection is useless if the reject gate occasionally misses.

Validate False Rejects, Escapes and Changeovers

Validation

Validation. Challenge the system with known defect sets and quantify false reject and escape rates across SKUs. Repeat after lighting, camera, bottle or software changes. For validation, define the defect or measurement decision first and then choose sensors, lighting, or logic that can prove it. If the symptom or performance target does not move as predicted, return validation to the baseline and investigate validation rather than stacking corrections. A vision system that is not challenged systematically can drift from the customer quality standard.

Questions that arise specifically in How to Automate ISBM Bottle Quality Inspection

Practical conclusion

A robust answer to How to Automate ISBM Bottle Quality Inspection should survive a restart and a full thermal stabilization period. The setup record should therefore connect defect library with base and gate and the bottle result from validation. A single top camera may miss a defect on a recessed base.