This website uses cookies primarily for visitor analytics. Certain pages will ask you to fill in contact details to receive additional information. On these pages you have the option of having the site log your details for future visits. Indicating you want the site to remember your details will place a cookie on your device. To view our full cookie policy, please click here. You can also view it at any time by going to our Contact Us page.

Send mistakes packing!

01 September 2020

Zohar Kantor explains how machine vision can offer an additional tool to help ensure that products are properly packaged and labelled. 

According to the World Allergy Organisation (WAO), 2.5%of the general population is allergic to one of more foods, while up to 10% can manifest adverse drug reactions. This is just one example of why in highly regulated industries, such as food manufacturing, it is particularly important to protect consumers by ensuring that items are properly packaged and labelled. 

Errors in packaging are often the cause of huge recalls — the wrong label on the wrong product, for example, could lead to food containing ingredients that it shouldn’t, with potentially dangerous consequences for consumers that have allergies. Other risks linked to incorrect packaging and labelling include premature spoiling, contamination with external pollutants and changes in the product’s taste and colour.

For these reasons, quality assurance (QA) needs to be an integral part of the packaging process. Yet the huge amount of stock keeping units (SKUs) and different labelling concepts required in the food industry to make this a particularly challenging task. 

Machine vision solutions can automate QA procedures, but their inherent complexity, coupled with the high number of SKUs and inspection scenarios, often prevents small and medium packaging plants from implementing them. 

Avoiding the cost of assistance 
Further, companies usually need the assistance of a systems integrator or machine vision expert to build and set up the solution. For many SMEs, the cost of these services, together with the initial investment in a traditional machine vision solution, are far too high. 

As a consequence, many small to medium sized packaging manufacturers rely on manual inspection, a method that carries an error rate of more than 25%. In the context of Industry 4.0, this is an anachronistic approach that cannot be relied on.

The introduction of Autonomous Machine Vision (AMV) by Inspekto provides users with universal inspection products that are completely autonomous – self-setting, self-learning and self-adjusting. This allows plant personnel to independently set up and operate the system, which comes ready to use, out of the box.

The INSPEKTO S70, has a number of characteristics that differentiate it from traditional QA solutions. Firstly, traditional solutions are tailor-made to inspect only one product at a specific junction of a production line and are designed and tested by a systems integrator to recognise all possible defects in a product.

In packaging, this means exposing the QA solution to thousands of defective items, until it memorises every potential production mistake, from missing labels, to misspelling, to unsealed or partially sealed packaging. This process of developing and training can take several weeks, causing extended periods of downtime.

The system can be installed in just 30 minutes by any plant employee while production is running as normal. The user needs to trace the perimeter of the area to be inspected and then present the system with an average of 20 to 30 good sample items, such as correctly printed and placed labels and integral packaging, and no defective items. The system, employs an array of AI engines working in tandem, to automatically learn what the end-product should look like and can then alert the QA manager if it detects any variation.

Checking seals
Autonomous Machine Vision systems are already being utilised in the Leicester plant of a multinational food, snack and beverage corporation, where the solution is being used to check that packaging is properly sealed and that labels are correctly attached and readable. The plant is planning to use the systems for other types of fast-moving consumer goods, to inspect the integrity of the packaging material and the correct placement of labels. 

Despite the fact that the solution cannot perform optical character recognition (OCR) or read text yet, it can compare the inspected label to images of a perfect label that it has previously memorised. This enables it to detect whether a label is incorrect, poorly printed, not securely attached, misplaced or missing altogether, and can also verify the physical integrity of the packaging material to make sure that the quality of products has not been compromised.  

Environmental changes, such as changing in lighting and product orientation, are not a problem for the system, which only flags up actual defects. The highly reflective properties of some packaging materials are therefore not an issue, because the system will self-adapt its camera parameters to ensure that a clear image of the product will be taken, with no human intervention. The system archives pictures of the inspected items, so that the manufacturer can respond to any claims and prove that items left the facility in perfect condition. 

Food allergies and drug intolerances are becoming increasingly publicised and regulatory compliance to labelling and packing guidelines is essential to guarantee consumers’ safety. Thanks to Autonomous Machine Vision, packaging manufacturers of every size now have an extra asset to improve their products’ quality.

Zohar Kantor is VP of sales and project management at Inspekto.


Contact Details and Archive...

Print this page | E-mail this page