More than meets the eye

To err may be human, but try explaining that to hordes of angry customers and consumers when a faulty food product slips through the system. Although manual inspections are an important part of quality control (QC) it is inevitable that mistakes will be made, which is a key reason why developments in analysis methods are high.

Equipment supplier Branscan has just launched Totalscan - an online QC system for pellets and granules, which uses a high sensitivity camera to transmit video data to the control station. "It takes images every 15-20 seconds of what's going through the conveyor," says sales and marketing co-ordinator Joanne Worthy-Jones. Totalscan works in real-time, so you can stop the machine immediately and check out any issues, she adds.

Objective opinions

"With visual inspections, three or four people could look at the same sample and find different problems, whereas our machinery is objective," says Worthy-Jones. "You set the specifications and the system will alert the manufacturer if there are any problems."

Dr Elfed Lewis, director of the Optical Fibre Research Centre at Limerick University, agrees that the ability of QC equipment to objectify results is having a positive effect on the food industry and believes it could set the foundations for tighter rules around product quality. "At the moment, people use manual operators, which can be a tiring task with subjective results," he says. "If you can quantify results, you can certify them, which will be a major issue in coming years with changing legislation."

A few years ago Lewis developed a colour sensor, which uses visible light reflection spectroscopy combined with an artificial neural network computer processing system to measure a food's colour (see Food Manufacture April 2005). Since then he has created a probe, which can measure both the colour, using optical-fibre sensors, and the temperature of cooked foods.

"We still use the same reflection spectroscopy technique, but we've introduced a more advanced artificial neural network and applied principal component analysis to make signal processing more efficient," says Lewis. He explains that the original network had many inputs feeding through numerous layers of nodes, but this has now been reduced to just three, making it much quicker.

Lewis is currently working with Echo Ovens, which provides equipment for the mass production of cooked products. "At the moment the probe is in the research and prototype stage at Echo's test facility, but the system has seen a lot of interest from the likes of McDonald's supplier Moy Park," he says.

At £20-40,000, retro-fitting the probe works out to be cost-effective when compared to the price of a new large industrial oven, which would set manufacturers back around £500,000, adds Lewis.

Cheap isn't so cheerful

But when it comes to food inspection, promoting a product's cost efficiency is not always the way to go. Flemming Møller, a scientist at Danisco, explains that when he initially applied a light scattering technique to measuring food particle size, colleagues were not convinced.

"The basic process is to shine a laser light on a sample and analyse the light scattering," he says. "Normally, we would have to use an X-ray, which is very expensive - around £200,000, but this method uses a laser pointer, which costs around £2 and a normal digital camera. The problem at first was that everyone thought it was very cheap, and therefore, it couldn't possibly be scientific."

Nevertheless, Møller continues to use the method, which is inspired by a technique used in the Lord of the Rings movies to make Gollum's skin look realistic.

He explains that adding emulsifier or hydrochloride to a product will affect the particle size, but that you would have to dilute a sample many times to analyse this in a traditional way, whereas his process gets instant results.

"We use this method when we are developing new recipes and evaluating new ingredients," he says. "I've used it for loads of applications: from bread with big bubbles; to milk where you're talking about microns."

While he admits that microscopy is still the most accurate way to measure particles, light scattering is more effective for multiple samples, which is why he believes it could be used in online quality control.

"The main drawback is that this is a new technique and in order to use the method, you need knowledge in a variety of areas: laser light, image analysis and particle size. But this a fast, cheap technology and I definitely see it becoming more widespread." FM