Automated Label Inspection Guide
Matt Wilton
Director
Table of contents
Share
Automated label inspection uses a camera and image-processing software to examine products as they move through production. Depending on the setup, it can check that a label is present and correctly placed, read printed codes, compare information with the approved product data, and reject a failing item. The challenge is choosing checks that work on the actual container, label and line conditions. Which checks do you need, and can they run on the line you already have?
1. What is automated label inspection?
Automated label inspection is an in-line machine vision check that verifies both the physical application of a label and the information printed on it.
The first half checks whether the label is present, correctly positioned and visibly intact. The second checks whether the printed content is readable and correct. A system may perform either half or both, depending on the production risk and available inspection position.
2. What does a label inspection system check?
The camera, lighting and software are selected to suit the defect and the product surface. One camera can perform several checks where the label is visible in a single image. A round container may need views from more than one direction.
Check | What it catches | Method |
|---|---|---|
Presence | Missing label or missing part of a label | Search for the expected label area, edge or printed feature |
Position and height | Label too high, low or off-centre | Measure label edges or reference features against set limits |
Skew | Label rotated relative to the container or pack | Compare label edges or graphics with the expected angle |
Double labels | Two labels applied in an overlap | Detect unexpected edges, thickness or duplicated print |
Creases, tears, bubbles and flagging | Visible damage, raised edges, trapped air or a loose label | Blob and edge tools, with lighting chosen to reveal the surface defect |
Wrong label against the approved master | Incorrect artwork, product variant or print layout | Pattern matching against an approved reference |
Date and lot codes readable and correct | Missing, blurred or incorrect variable data | OCR or OCV, with the expected code or format checked |
Barcode read, graded and cross-checked against printed data | Unreadable code or a mismatch between barcode and printed lot or product data | Barcode decoding, quality grading and data comparison |
Print quality | Missing characters, smudges, breaks or inconsistent print | Contrast, edge or pattern analysis; OCR/OCV where text is involved |
A successful read does not by itself confirm that a code belongs to the correct product. For example, the system can decode a barcode, read the printed lot code and compare selected fields. The expected values must come from a recipe, production order or other defined source.
Read more about Barcode and 2D code reading here.
3. How to get reliable inspection on difficult labels?
Reliable label inspection starts with specifying the camera, lens and lighting for the actual product and defects. AIET Group’s application experience is key: a setup that works for a matte paper label may not produce a stable image on a clear label, glossy film, glass bottle or curved PET container.
The lighting is chosen to make the required features visible. Diffuse illumination can even out reflections; low-angle light can reveal raised edges, creases or bubbles. Polarising filters can suppress glare from reflective surfaces, while selecting a suitable wavelength can increase contrast between the label, print and container. Shrouding around the camera and inspection area helps control changing ambient light.
For round containers, AIET Group specifies multiple camera views or a 360° optical arrangement to cover the required surface. Where products move quickly, strobed lighting and encoder triggering help capture sharp, consistently positioned images. The setup is assessed on sample products so the inspection tools are built around the image the system can reliably produce.
4. How fast can label inspection run?
Inspection speed depends on the image, the number of checks, the number of views and the time available before the product reaches the reject point. Presence and position checks generally need less processing than reading and validating several fields of variable text.
As a planning guide, intermittent carton and pouch lines may run from tens into the low hundreds of units per minute. Continuous bottle lines can run at up to 600 products per minute.
At higher speeds, a short exposure helps prevent motion blur. A strobe can deliver a brief, bright flash during image capture. An encoder trigger can synchronise acquisition with product movement, which is useful when conveyor speed varies. The inspection and reject timing must be tested together at the intended line speed.
5. Which label inspection setup fits your line?
AIET Group offers three routes for machine vision label inspection, depending on who will engineer the controls and install the equipment.
Route | What’s supplied | Who integrates | Typical installation time |
|---|---|---|---|
Retrofit camera | Camera, lens and lighting. The scope can be hardware supply only or include integration. | The customer’s controls engineers integrate it if they have the capability. If not, AIET Group can handle the integration, PLC handshake and commissioning. | 1 hour for hardware installation. 1–2 days if AIET Group also handles integration and commissioning. |
Turnkey retrofit | Camera, lighting, mounting, reject mechanism, PLC handshake and commissioning on the existing line. | AIET Group delivers the working system. | 1–3 days on site. |
Machine builder integration | Inspection system and application support during the machine build. | The machine builder can integrate it with AIET Group support, or AIET Group can complete the integration. | 1–3 days for inspection integration and commissioning, scheduled within the machine build. |
A camera-only supply requires the customer to handle mounting, triggering, reject logic and production testing. A turnkey retrofit includes those elements on the existing line and is handed over working. For a new machine, bringing the inspection application into the build allows the camera position, lighting and product handling to be considered together.
6. What happens when a label fails?
The inspection system sends a pass or fail result to the line controls. A PLC tracks the product to a reject point, where a mechanism such as an air blast or pusher removes the failed unit. The reject method depends on product weight, conveyor layout, spacing and the time between camera and reject station.
Image and result logging can associate a captured image and inspection data with a production event for traceability or fault review. At product changeover, an authorised operator can select or teach the new product recipe, including the expected label pattern, code format and inspection limits. The new recipe should be checked against known good and deliberately faulty samples before production resumes.
7. What to prepare before you call us
Container types and sizes, including the range handled on the line.
Label material, colour and finish, including clear or metallic areas.
Line speed in units per minute and any speed changes during a run.
Conveyor photos showing the space either side of the labeller and possible camera positions.
Codes printed on the label, including the date, lot code and barcode type.
Existing reject mechanism and PLC, including how a reject signal is currently handled.
Current defects getting through, with sample products or clear photographs if available.
How AIET helps
AIET Group reviews sample products and labels to assess the inspection task, then uses lighting trials to check whether the required features can be seen reliably. An application review considers camera position, lens, lighting, inspection tools, line signals and the reject point. The project can proceed as a camera, lens and lighting supply for an in-house controls team, a turnkey retrofit, or an application integrated into a new machine build. AIET Group provides local support across the UAE and GCC. See the machine vision label inspection service.
Speak to AIET Group
To scope an application, send sample labels or containers, the line speed, photos of the conveyor and labeller, and a description of the defects getting through. Include the existing reject arrangement and PLC details if available. This information helps determine whether a camera can cover the required checks from one position or whether the application needs multiple views. Contact AIET Group about a label inspection application.
Frequently asked questions
What is automated label inspection?
Automated label inspection uses a camera and image-processing software to check labels on products as they move through a production line. Depending on the application, it can verify label presence and position, inspect visible defects, read printed codes, compare data and signal a reject for failed products.
Can label inspection be added to an existing production line?
Yes, label inspection can be retrofitted to an existing production line if there is a suitable camera position and enough time to trigger the inspection and reject a failed product. The retrofit must account for product handling, lighting, PLC signals, mounting space and the reject mechanism.
Can a camera check the expiry date is correct, not just printed?
Yes, a camera can read an expiry date with OCR or OCV and compare the result with an expected format or value supplied by the production system or recipe. The source of the expected date must be defined; detecting printed characters alone does not confirm they are correct.
Does label inspection work on clear labels?
Yes, clear labels can be inspected, but the lighting and camera view must reveal features that distinguish the label from the container. Reflections, background colour and container curvature affect the image, so sample testing is needed to confirm which checks are reliable.
How fast can a label inspection system run?
The achievable speed depends on the checks, image quality, product spacing, number of camera views and reject timing. A presence check usually needs less processing than OCR combined with barcode verification. The complete application must be tested at the intended line speed.
What is the difference between OCR and OCV?
OCR, optical character recognition, reads characters and returns them as data. OCV, optical character verification, checks characters against an expected value or reference. A label system can use OCR to read a date, then compare it with the required date or format. Read more here.
How long does it take to set up a new product?
Setup time depends on how different the new container, label, print and inspection criteria are from the existing recipe. A change may need new image regions, code formats, reference patterns and limits. The recipe should be tested with known good and faulty samples before release.
Can label inspection be built into a new machine?
Yes, label inspection can be integrated into a new labeller, filler or packaging machine during the build. The machine builder can plan camera mounting, lighting, product triggering, PLC communication and reject handling around the product flow, with application support during development.
