sweets and confectionery as, for instance, plastic particles, chips, threads, filaments or insects, spiders, or other small animals. These kinds of particles can be easily detected with colour images as long as they’re in a position where they are visible,” he says. I hope I am not the only one who positively shuddered
at the thought of a spider in my chocolate bar, but the real pressing question I had was what if the foreign particles are of the same colour/hue as the product? For instance, white plastic particles in white chocolate or white cream? “Yes, this brings us to the second solution digital image processing can offer,” Stephan laughs. “Advanced colour image processing is where wavelengths outside the visible spectrum are analysed using, for instance, infrared cameras. Another alternative is analysing complex combinations of different colour spectra using multi or hyperspectral cameras. Using these technologies, foreign particles of the same colour as the product can be detected, such as white plastic in white chocolate or brown nutshells among brown nuts. Additionally, mould infestation and rotten spots can be detected,” he explains. “Another very up-to-date alternative is the analysis of elaborate patterns using complex algorithms, including deep learning trained neural networks. Using this technology, things such as shell-nut-differentiation are possible but also judging and evaluating the overall impression a product makes, Stephan says. Consumers make buying decisions based on a number of factors, one of which is how visually appealing a product is. Stephan suggests this appeal is difficult to quantify in measurable data, and it actually goes beyond factors such as colour, shape, flawlessness. “Judgement might be based on colour and shininess, colour gradients, etc. Customers may prefer an appearance that is maybe more rustic, with a ‘hand-made-impression’ in one product, perfect, smooth, and symmetrical in another. It might be based on the distribution of the topping – sugar coating, sugar pearls, or other decorative elements. Since all of these details cannot be taught to the detection system individually, deep learning is used.”
“A trained and well-versed quality management employee evaluates a number of products according to their ‘beauty’ (all kinds of scales are possible here – such as ‘good – ok – bad’, school grades, quality classed, or simply IO-NIO) and enters his/her judgement into the system. The system learns and subsequently is able to form its own judgments on the beauty or desirability of the product,” he says. It is noteworthy that not only the appearance and quality of the end product can be judged but also the appearance and quality of the raw materials such as cocoa beans.
How has technology advanced weighing and detection processes?
Detection techniques and technologies have improved in a number of ways. “Weighing systems such as checkweighers can perform more accurate measurement at high speeds through a new generation of load cells; x-ray and metal detection systems are capable of increasingly sensitive detection through improved software algorithms; detection systems are becoming smaller and more affordable, and can be integrated with additional detection capabilities through the advent of ‘combination’ systems, which can, for example, match checkweighing with metal detection in a single unit,” explains Daniela Verhaeg.
“The system learns and subsequently is able to form its
own judgments on the beauty or desirability of the product”
There has also been a move towards more automation. Manual processes are known to be slow, labour-intensive and prone to human error. Efficient manufacturing is important in all sectors, and automation enables faster, more accurate and more efficient production. The nature of product inspection today is characterised by high levels of automation. “Sorting technologies achieve an accuracy that manual
TOMRA Nimbus BSI 28 Kennedy’s Confection April 2022
sorting simply cannot. And at the same time as looking over the production line like guardian angels, automated sorters also enhance product hygiene, solve labour-related challenges, increase throughput, maximise yield, and gather data that can unlock further improvements in line efficiency,” says Christian Hofsommer. He adds: “In addition to taking care of food safety and product quality, sorting machines also help solve the challenges traditionally associated with employing manual sorters - an effective pill for headaches caused by labour scarcity, cost, variable effectiveness, and absenteeism. And whereas manual sorting is unavoidably subjective, imperfect, and more vulnerable to error when labourers are tired or bored, automated sorters can work for hour after hour with superior accuracy, consistent standards, and unflagging efficiency. “So yes, for these reasons, there is a move towards automated sorting, and this trend will only strengthen in the future. At the same time that consumers are getting pickier about food quality standards, it is getting more difficult in many parts of the world for confectionery producers to recruit and retain workers for tasks such as sorting.”
KennedysConfection.com
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44
