KOMPASS adds lot code verification to every pack on the line. AI reads each lot, batch and date code, compares it with the batch record for the running order and signals the line when a code is missing, unreadable or carried over from the last run.
Lot and batch codes link every pack to how it was made. When the code is missing, illegible or belongs to the previous order, that link breaks, and a recall can widen to stock that was never affected.
Our AI Product Inspection system, KOMPASS™, handles lot code verification with models trained on your print, checking each lot, batch and date code against the active order. Results reach the line controls with an image of every failure.
Jidoka's demonstration shows lot code verification in real time: KOMPASS reads batch codes, dates and serial numbers as packs pass the coder. A print-synchronised trigger captures each pack, and inspection results and NG images are recorded for digital traceability.
A trigger linked to the coder captures every pack with the code in view. Controlled lighting isolates the printed region on the substrate.
The model reads each field and checks it against the active order, so wrong, missing and unreadable codes are all caught.
Failing packs are rejected or held, repeated failures alert the team, and every result is stored with its image.
Send us examples of good, faint and wrong codes and how your order data reaches the coder. We will assess the imaging and integration the check needs.
Get In TouchEach method needs a readable code. The difference is whether lot code verification checks the content, and how quickly a coder problem reaches the team.
Appasamy needed one inspection to cover lens presence, printed codes and labels, QR-based serial number consistency, particles and packaging integrity across 4 product variants and 8 lens shapes. KOMPASS sorts every pack OK or NG.
These are published results for the complete lens packaging inspection, of which code checks are one part.
Lot code verification checks that the code says the right thing, not only that it is legible. KOMPASS reads each field and compares it with the batch record or active order. A pack fails if the code is missing, unreadable or different from what should be printed. That catches the most expensive error, a perfectly printed code that belongs to another batch or another day. Presence sensors and simple readers cannot make that comparison, because they do not know which code is correct.
Usually from the same source that feeds the coder: the work order, a recipe on the line controller or a manufacturing system. When the order changes, the expected code changes with it, so operators do not have to type values into the inspection station. Where that link does not exist yet, a supervisor can confirm the code at the start of each run. The link is tested during commissioning by running a planned changeover and confirming the reference updates.
Yes. The switch between orders is when most coding errors happen, because the coder message is updated late or not at all. Because every pack is compared with the new order, the first pack with an old lot number fails immediately. The station can also stop the line after a set number of consecutive failures, so a whole pallet is not coded wrongly before anyone reacts.
Missing codes usually come from a printer skip, a jam or a product that reached the coder out of position. Smudged print comes from contact before the ink dries, low ink, a blocked nozzle on the inkjet coder or condensation on the pack. The failure images stored by the station show which pattern you have, which points maintenance to the right fix. Trends by coder and shift show whether failures are random or tied to one machine.
Yes. Date code verification uses the same reading, with a rule for what the date should be. An expiry date is usually calculated from the production date and shelf life, so the station can check both the format and the value. Best-before, use-by and manufacturing dates can be checked together with lot fields when they share the same print area. Where packs carry several dates, each field is checked against its own rule.
Every inspection result is stored with the code read, the expected value and a time stamp, and NG images are kept for every failure. That creates a digital record that each pack was checked, which supports traceability and is the evidence auditors ask for when they review batch code verification. If a recall happens, the record helps confirm which codes actually left the plant.
A print trigger from the coder or a product sensor tells the camera when the code is in position. That keeps the code in the same place in every image and avoids reading a pack before printing has finished. Reading on fast lines depends on that timing as much as on reading, so it is set up and tested during commissioning. If products arrive unevenly spaced, the trigger also prevents two packs being captured in one image.
Send samples of good, faint, smudged and wrong codes, the code format for each product and how order data reaches your coder today. Add the line speed, pack materials and the reject position. With that, we can recommend camera placement, lighting and integration for lot code verification on your line and plan the validation.
Share your code samples, formats, order data source and line speed. We will plan lot code verification at your coder and give you a timeline and budget. Expect go-live in roughly 6 to 8 weeks.
