Vision AI · MI-VAI-03
Label OCR & Piece Counting
Reads serials/batch codes on a moving line, counts pieces on a tray, and pushes results to a sheet or ERP endpoint.
Vision AI · MI-VAI-03
A fixed-focus camera with a bar light and an edge AI board that reads serial and batch codes on a moving line, counts pieces on a tray, and posts the result to a sheet or your ERP.
- ₹28,000pilot BOM · hardware only
- 4 – 6 daysto a live pilot
- Intermediateintegration level
- 3validated providers
Why hand-keyed serials break traceability
The problem
Serial numbers and batch codes are keyed in by hand at packing and dispatch. Carton counts are recounted because nobody trusts the first count. A wrong digit means a traceability gap; a wrong count means a customer debit note.
- Manual data entry errors break traceability
- Recounting at dispatch costs time and still misses
- No link between the physical label and the ERP record
- Audit trails depend on paper and memory
The solution
A camera with a fixed-focus lens and an LED bar light captures each label or tray; an edge AI board runs OCR for the serial and batch code and a counting model for pieces. Results are validated against expected formats and pushed over REST to a sheet, WMS or ERP endpoint, with the image kept for audit.
- Reads printed and dot-peen serial and batch codes
- Counts pieces on a tray or in a carton in one shot
- Validates format and flags unreadable labels for a human
- Posts to ERP/WMS via REST with the image attached
Validated service providers
Validated by TwoElectrons and ranked by validation score: verified credentials, completed jobs on the platform and client ratings.
- Sample Provider J — Bengaluru · Inspection & defect models · ★ 4.9 · 15 jobs
- Sample Provider K — Noida · Video analytics & PPE · ★ 4.7 · 9 jobs
- Sample Provider L — Pune · OCR & counting systems · ★ 4.6 · 8 jobs
Provider listings are samples pending live registrations.
Pilot BOM and cost
Reference hardware to run the pilot. Unit costs are indicative Indian market prices for the component class.
| Qty | Item | Unit cost | Line |
|---|---|---|---|
| 1× | Edge AI compute board | ₹18,000 | ₹18,000 |
| 1× | Camera module with fixed-focus lens | ₹4,500 | ₹4,500 |
| 1× | LED bar light | ₹3,000 | ₹3,000 |
| 1× | OCR + counting model, REST output | Free tier / open source | — |
| 1× | Mount & enclosure | ₹2,500 | ₹2,500 |
| Pilot BOM total | ₹28,000 |
POC BOM cost only — hardware for a pilot. Installation, integration and provider services are quoted separately. Request for Quote
How camera OCR and counting work
- Mount. Fix the camera and bar light over the label or tray position.
- Calibrate. Set focus, exposure and the read region.
- Validate. Run OCR against 200 known labels; tune for your fonts and printers.
- Connect. Post results to a sheet or ERP endpoint.
Pilot architecture
Camera + LED bar light (fixed-focus lens) → Edge AI board (OCR + counting model) → REST endpoint (sheet · WMS · ERP) → Operator screen (read · count · exceptions)
OCR and counting run on the edge board; results and images are posted to your system of record.
Business case: errors, recounts, traceability
Ranges are typical figures reported for this class of solution; your pilot establishes the numbers for your plant.
- > 99% read accuracy on printed codes (after tuning for your printers)
- Seconds per tray count (versus manual recount)
- Zero hand-keyed serials (at the covered station)
- 4 – 6 days to a live pilot (one station)
Who this is for: Dispatch and logistics heads, Quality and traceability leads, Auto components (IATF traceability), Pharma and FMCG packing, Electronics and appliance assembly, Any plant keying labels by hand.
Illustrative case study
Illustrative scenario · not a client reference
A bearing plant's packing line with per-box serials in Jaipur (illustrative)
Set-up. Camera and bar light over the label station, edge board under the bench, results posted to a shared sheet and then to SAP via the plant's middleware.
What happened. Read accuracy reached 99.6% after tuning for the thermal-transfer printer; the operator only handled the flagged 0.4%. Tray counts were accepted at dispatch without recount, and a customer traceability query was answered from the image archive in minutes.
- Pilot budget: ₹28,000 hardware
- Read accuracy: 99.6%
- Recounts at dispatch: eliminated
FAQ
Which code types can it read?
Printed alphanumerics, 1D/2D barcodes and QR codes, and — with tuning — dot-peen or laser-etched marks on metal.
What happens when a label is unreadable?
It is flagged with the image for an operator to key manually; the exception rate is tracked so printer issues are caught early.
Can it talk to SAP or our WMS?
Results are posted over REST; most plants connect via existing middleware or a simple endpoint in the scale-up phase.
What does the ₹28,000 cover?
Camera, lens, bar light, edge board and mounting. OCR tuning, ERP integration and provider services are quoted after a Request for Quote.
