Product inspection: Breaking down barriers

26 May 2022

John Dunlop argues the case for inspection-as-a-service in the food production sector.

Continue reading this article

Register now for free and access every article and to register for the print edition.

With the UK labour force facing sharp shortages and deep-learning now being served up in bite-sized portions it is time for even the most cautious food processing company to ask whether the benefits of automated inspections outweigh the inherent risks to operational output.

There are some myths surrounding automation that need to be discarded as they are simply passed their ‘sell by’ date. Take, for example, the idea that automating processes is just for the ‘big boys’. As with any emerging market, the high initial cost meant that early adopters were likely to be larger companies with deeper pockets, but this is no longer the case; the demographic is growing while the costs across the board are shrinking.

Let’s start with one of the key ingredients – cost. Kit that might have seemed prohibitively expensive just a few years ago is already more affordable due to a deluge of demand. A piece of advanced inspection technology that may have cost upwards of £16,000 just a few years ago can now be brought in for as little as £5,000.

Then, there’s the issue of ebbing employment. In the UK alone, there are currently 1.3 million vacancies – which is up by 50,000 from last quarter, according to the latest labour market statistics. With over 200,000 more younger people now absent from the workforce and many of the older demographics choosing to retire amid Covid concerns, over half of businesses aren’t able to meet the demand for labour.

Without immediate intervention, employers are looking at the very real possibility of a deserted factory floor, and automated inspections are one way to ensure production can continue.

Cost versus quality

There is some concern in the sector that the traditional accept/reject model of automated inspection simply isn’t suited for food processing – but this is not always the case. It is not ideal for every application, but it is increasingly suitable for the majority.

Take a bakery line, for example. The bulk of quality control for these are unautomated - relying on manual visual approval, based on what a trained operative has been taught to observe.

Inspection-as-a-service has come on in leaps and bounds in just a few short years and many of the more adept models can be trained, like a new staff member, to detect key factors which may once have been considered out of their remit.

From the perfect crispy bronze finish of a pastry to the right rise on a croissant and even the ideal density of a doughnut, the established technology can be tuned to detect all the visual indicators of clear quality control that a manual operative can - and many that they can’t.

In this sense, it yields all the benefits of subjectivity but with a constant, unerring baseline specifically tailored to meet pre-set parameters.

While automating production can still take time, inspection-as-a-service is able to be rapidly deployed, because unlike building a bespoke outfit, it can be set up in as little as 48 hours and offers a temporary solution to quality concerns that can yield immediate results.

Bytronic, for example, has recently introduced a South-East Asian food multinational to inspection-as-a-service to curb excessive waste and knock-on production line costs caused by mandatory downtime due to contamination. The factory, which produces soy sauce, puts out around 3 pouches every second across each of its seven production lines. Inspection-as-a-service helped prevent problems and drive down costs.

While there are still barriers to inspection-as-a-service, one of the most basic ones which must first be overcome is a reluctance to upset established production lines – even when presented with productivity gains.

John Dunlop is chief technology officer and founder at Bytronic Vision Automation.