Why "The Scanner Read It" Is a Low Bar
Ask most line supervisors how they know their batch codes and barcodes are correct, and the answer is usually some version of: "the scanner beeps, so it's fine." That confirms the code was decodable at that exact moment, on that exact scanner, under that exact lighting. It says nothing about whether the printed date matches what was actually produced, whether the print quality will still be readable after a few weeks in a warehouse, or whether the code was even legible on the far side of the container that nobody scanned.
Machine vision closes this gap because it does two genuinely different jobs that are often conflated as one: reading the code (can the data be decoded) and verifying it (is the print quality good enough, and does the decoded data match what should be there). Most inspection failures on real lines trace back to only one of those two jobs being done.
A recall or customer complaint over an unreadable or incorrect expiry date almost never happens because the printer produced garbage on every unit — it happens because the printer produced garbage on 40 units out of 40,000, right in the middle of a shift, and nobody was checking every single one. Sampling catches trends. It does not catch the one bad run of 40.
What Machine Vision Actually Checks on a Code
Barcode, batch code, and date code inspection on a packaging line covers several distinct checks, and a serious system does all of them — not just the first one.
Presence & Position
- Barcode/QR code is present at all
- Code is positioned within the expected zone
- Code is not rotated, cropped, or partially printed off the label edge
- Correct code type applied (right SKU's barcode, not a leftover from the previous run)
OCR / OCV — Batch & Date Text
- Batch/lot number characters are legible and complete
- Printed date matches the expected manufacture/expiry string for that run
- No smudging, missing dots, or partial character strikes
- Correct date format for the target market (DD/MM/YYYY vs MM/DD/YYYY)
Print Quality Grading
- Bar width and spacing within tolerance
- Contrast between bars/text and background substrate
- Edge definition — no bleed, no broken bars
- Consistency across the full print run, not just the first unit checked
Data Cross-Check
- Decoded barcode data matches the batch/date printed alongside it
- Both match the expected values from ERP/MES/PLC for the active run
- Serialisation sequence has no gaps or repeats, where serialisation applies
- Timestamped accept/reject decision logged per unit for traceability
Barcode Reading vs Barcode Verification: The Distinction That Matters
These two terms get used interchangeably in casual conversation on the shop floor, and the difference is exactly where most gaps hide.
| Aspect | Barcode Reading | Barcode Verification (OCV / Grading) |
|---|---|---|
| What it confirms | The data can be decoded right now, on this camera | The print quality meets a defined grade, so it will decode reliably on any scanner, for the life of the product |
| Catches print degradation trend | No — only fails once unreadable | Yes — flags declining contrast/edge quality before it becomes unreadable |
| Confirms correct data (not just readable data) | Only if compared against expected value | Yes, when combined with a data cross-check step |
| Typical failure it misses | Marginal print quality that reads today, fails at the retailer in 8 weeks | — |
| Where it belongs | Minimum baseline — every line should have this | Recommended wherever code failure has regulatory, recall, or brand consequences |
Not sure if your current setup verifies codes or just reads them?
Send us your current code inspection setup and a sample label. Our applications team will tell you plainly which gap you have — presence, print quality, or data correctness — before recommending anything.
Why Printers Drift and Vision Catches It
No printer holds perfect print quality indefinitely through a production shift, and this is not a defect in the printer — it is normal wear behaviour that inspection exists to catch.
- Thermal transfer ribbons wear unevenly across the print head width, producing faint bars on one side of the code well before the ribbon is fully spent
- Ink-jet nozzles partially clog over hours of operation, dropping individual dots from characters — often just enough to make a "3" look like an "8" to a human eye but fail an OCR confidence check
- Print head pressure on contact printers drifts with temperature and mechanical wear, changing bar width tolerance gradually
- Substrate variation — glossy film, curved surfaces, metallic backgrounds, or low-contrast label stock — makes the same printer settings produce different real-world print quality across SKUs
All four of these degrade gradually. A printer producing perfect codes at 8 AM can be producing marginal codes by 2 PM with no alarm on the printer itself, because the printer has no way to know its own output quality — only a camera looking at the finished print does.
They install a camera, confirm it reads the barcode on day one, and consider code inspection solved. What actually happened is they validated decodability under ideal conditions on a good print sample — not print quality grading, not data cross-checking against the ERP recipe, and not ongoing verification as the printer drifts through a shift. The camera was capable of all three from day one; it was simply configured to do the easiest one.
Implementing Code Verification: What to Get Right
1. Decide What You're Actually Protecting Against
A missing barcode, an unreadable batch code, and a correct-looking-but-wrong expiry date are three different failure modes with three different consequences — a rejected shipment, a warehouse-level recall, and a consumer-facing compliance breach, respectively. Design the inspection scope around the failure mode that matters most for your product category, not around whatever the vendor's standard package includes.
2. Connect to the Line's Actual Production Data
Data cross-checking only works if the vision system knows what the correct batch number and date should be for the current run — pulled from ERP, MES, or the printer's own PLC recipe. A system checking print quality in isolation, with no connection to expected data, cannot catch a correctly-printed wrong date.
3. Set the Print Quality Threshold Deliberately
Grading standards for barcode quality exist in the industry (commonly referenced as ISO/IEC 15416 for linear barcodes) and give a structured way to define "good enough" rather than relying on a human eye's subjective judgement. Agree the acceptance grade with your quality team before commissioning — set it too loose and marginal prints slip through; set it too tight and you reject codes that would have scanned perfectly well at every point in the supply chain.
4. Keep a Timestamped, Exportable Audit Trail
Every accept/reject decision should be logged with a timestamp and, where practical, linked to the batch or serial number of the unit — not just a pass/fail counter. This is the record that answers "which units were affected" when a printer drift event is discovered after the fact, rather than "we're not sure, check the whole batch."
Industries Where This Matters Most
- Pharmaceuticals: Batch and expiry legibility tied directly to patient safety and regulatory labelling requirements; barcode integrity supports track-and-trace and anti-counterfeiting programmes
- FMCG & Food and Beverage: Expiry date accuracy drives consumer trust and retailer compliance; incorrect dates are a leading cause of retailer rejection and chargebacks
- Nutraceuticals & Personal Care: Growing export volumes into markets with strict GS1 and labelling enforcement
- Agrochemicals: Batch traceability requirements tied to regulatory approval and field-safety recall capability
Practical Takeaway
"The scanner beeped" tells you a code was decodable at one moment, under one set of conditions. It does not tell you the print quality will survive the supply chain, and it does not tell you the batch number and date are actually correct for what left the factory that day.
Machine vision that combines presence detection, OCR/OCV against expected data, and print quality grading — with a timestamped audit trail — is what actually closes that gap. It is not a bigger version of the same check; it is three different checks that most lines only run one of.
Ready to Close the Gap on Code Verification?
Optomech's label inspection systems verify barcode presence, print quality, and batch/date code correctness inline, at up to 14,000 units per hour, with full audit-trail export. We've supplied pharma and FMCG packaging lines across India for over 40 years.
Frequently Asked Questions
What is the difference between barcode reading and barcode verification? ▾
Barcode reading confirms that a scanner or camera can decode the data in the barcode. Barcode verification checks the print quality of the barcode itself — bar width, contrast, edge definition — against a grading standard, so that the code remains readable reliably by every scanner downstream, including at a retailer's point of sale months later when the print may have degraded further. A line can pass "reading" every single unit and still ship codes that fail at the shelf if only presence and decodability were checked, not print quality.
Can machine vision verify that a batch code and expiry date are correct, not just present? ▾
Yes. This requires OCR (optical character recognition) or OCV (optical character verification) combined with a data comparison step. The vision system reads the printed characters and compares them against the expected batch number and date string — typically pulled from the line's ERP/MES or PLC recipe for that production run — and flags any mismatch, missing character, or smudged print that falls below a legibility threshold. Presence-only checks do not catch a wrong date or transposed digit; OCR/OCV comparison does.
At what speed can machine vision inspect barcodes and batch codes on a packaging line? ▾
Speed depends on the packaging format and inspection scope. Optomech's label inspection systems (LIS), which include barcode and batch/date code verification, run at up to 14,000 bottles per hour on 2-sided flat-bottle inspection and 5,000 containers per hour on in-mould label (IML) inspection, with a 360-degree variant for round bottles where the code location varies with label wrap. Inspection is synchronised to line speed so no container passes uninspected.
Why do barcode and date code print quality issues happen even when the printer is working? ▾
Thermal transfer ribbons wear unevenly, ink-jet nozzles partially clog and produce faint or missing dots, print heads drift out of pressure calibration over a shift, and substrate variation all degrade print quality gradually rather than suddenly. A printer producing perfect codes at shift start can be producing marginal or unreadable codes by hour six without any alarm firing on the printer itself — which is why inline vision verification, not printer self-diagnostics, is the control that actually catches it.
Is machine vision code verification required for pharmaceutical and FMCG compliance in India? ▾
Regulatory frameworks increasingly expect documented control over batch and expiry code legibility — correct, legible batch and expiry information is a basic requirement under India's drug labelling rules and FSSAI packaging norms, and GS1 barcode standards are widely adopted for traceability and anti-counterfeiting programmes. While the specific validation approach depends on your product category and export markets, 100% automated inline verification with a timestamped, exportable audit trail is the practical way to demonstrate control during a customer or regulatory audit.