Label inspection quality control

Improving quality control for margarine maker with Machine Vision label inspection system

margarine mixer
  • Fully automated visual product inspection system for one of the world’s largest producers

 

  • Correctly matches multilingual lids and tubs, reducing operating costs and improving quality detection rates

 

  • Project rolled out across 50 production lines over six sites

 

  • Supporting sustainability targets and protecting reputation for quality.

Delivering marginal gains for almost £200m worth of spreads & butters every year

The Challenge

We recently worked with one of the world’s largest producers of margarine, responsible for millions of pounds of spreads and preserves sold across the UK and Europe.

When it needed an automated way of ensuring every product for each country, variant and promotion was correctly labelled and packed before dispatch, its longstanding supply-chain partner Bytronic delivered the solution.

margarine maker

Split-second label checks using advanced machine learning

Our Solution

The large-scale project required quality control systems that could be rolled out to more than 50 interconnected production lines over multiple sites.

Previously, these products were checked by manual inspections; a slow, costly and inaccurate approach.

Because of the nature of the products, it was vital that labelling on allergens, health and safety and consumer rights for each product and region was accurate, without fail. 

Manual inspections or less sophisticated systems could easily miss subtle differences between products, such as language differences or incomplete or inaccurate packing.

Due to the nature of the production process – with components loaded in stacks that obscure the labelling – the tubs and lids could not be checked before assembly without adding costly delays. 

Instead, they must be checked in the final moments before the lids and tubs are combined on the line.

Retrofitting 50 production lines over multiple separate sites

When designing our solution, it was important that we delivered the fastest, most reliable and cost-effective way to spot labelling issues, in an automated system. One option – using artwork codes printed on each component – is often small and tricky to read. Barcodes can’t be relied on either, because often seasonal product variations use the same codes.

We needed to deliver an automated solution that would overcome these challenges, checking every pack before assembly, without causing delays. So we designed and installed a series of bespoke machine vision systems to carry out instant visual inspections of every component, milliseconds before the lids are applied to the tubs.

By combining high-resolution optical cameras, advanced pattern detection, image recognition software and programming, our systems identify the graphics on each tub in a little as 100ms, using machine learning to identify that every lid and tub are a match in the split-second before assembly.

If a mismatch is detected, the incoming lid is quickly rejected and replaced with the correct design.

Makes product mismatches or shipping errors almost impossible

The Result

Now up-and-running in 50 lines across multiple sites in different countries, the system has made product mismatches or shipping errors almost impossible – reducing the need for returns or recalls.

It has also made a significant impact on waste reduction, supporting the client’s sustainability ambitions.

We’re proud to be helping one of the world’s largest producers of spreads and margarines. Our system helps the company meet its legal, safety and sustainability requirements and protects its brands’ reputations for quality.

Find out more about our label inspection solutions here.