Speed comes first on a bottling line. A filler running 600 bottles a minute gives each bottle 100 milliseconds in front of a camera. Every check has to fit inside that window, and the reject gate has to act on the right bottle afterward. Speed shapes everything else.
How Is Fill Level Checked?
Fill level gets checked first because it carries legal weight. In India, the Legal Metrology (Packaged Commodities) Rules, 2011 set how far below the declared quantity a pack may fall. US inspectors check net contents against the procedures in NIST Handbook 133. An underfilled batch can mean a fine and a market recall.
On clear and translucent bottles, a camera with a backlight finds the liquid line and measures its height against a target band. Foam on carbonated drinks and milk products blurs that line, so the camera sits further down the line where the foam has settled. Opaque bottles, cans and cartons hide the liquid completely. Those lines use X-ray or high-frequency sensors, which read through the wall. Glass is the easy case.
Overfill matters too. On a line filling a million bottles a month, every extra milliliter is product given away.
What Does a Cap Check Look For?
The cap check starts with presence. After that, a camera looks at height and angle to find a cocked cap, which sits high on one side because it crossed a thread. A cocked cap can leak, and it often passes a quick look from above.
Color and type come next. A line that runs several flavors with different cap colors needs to confirm that the right cap went on the right product, especially after a changeover.
No camera can see torque. Quality teams measure it offline with a torque tester, on bottles pulled from the line every hour or so. Vision catches the visible signs of a bad application, and the torque test confirms the force.
Why Does the Tamper Band Need Its Own Check?
A tamper band is the ring below the cap that breaks on first opening. Consumers and retailers read a broken or missing band as a sign the product was opened, and they pull it from the shelf. Bands fail in the capper. The bridges that hold the band to the cap can snap if the capping head applies too much downward force, or the band can fold up under the cap. A camera looking at the neck from two or more sides finds broken bridges and lifted bands. One camera from one side misses half the ring.
Which Label Faults Matter Most?
Label checks on bottles cover presence, position and skew, plus the content itself. Round bottles add a challenge. The label wraps around the curve, so a single camera sees only part of it. Lines often use three or four cameras or a mirror setup to see the full wrap.
Content checks confirm the right label went on the right product and that the batch code and expiry date printed correctly. A wrong label after a changeover is a common recall trigger in food and drink. It is also the easiest one to prevent, since the camera only needs to compare the label with the production order.
How Do You Put the Checks Together?
Good bottle inspection systems combine these checks at two or three points on the line. Fill and cap usually share a station right after the capper, with the tamper band and label checks further downstream.
The reject system deserves the same attention as the cameras. A pusher or air blast must hit the exact failed bottle at full speed. A confirmation sensor after the reject gate should then verify that the bottle actually left the line. Without that sensor, a jammed pusher lets every failed bottle through while the screen shows rejects. Nobody notices until a customer does.
Start with a short audit. Pull the customer complaints and internal holds from the last year and sort them by fault type. If underfill and cocked caps top the list, a vision cap and fill inspection station after the capper is the first purchase. Label errors at the top of the list point downstream.
Then test at full line speed with real defects. Pull bottles with low fills, cocked caps and broken bands from quarantine stock and run them through mixed with good bottles. Bottle inspection systems that reject every planted defect at full speed, and log the image for each, are ready for production.
A vision cap and fill inspection station that misses even one planted bottle needs tuning before the go-live date.
FMCG teams can get in touch with Jidoka Technologies to plan vision AI checks for fill levels and caps.