In recent year, Artificial Intelligence (AI) has become a buzzword in almost every industry. In this article, we delve into the topic of AI in optical sorting with Stuart Bashford, Head of R&D at Bühler SORTEX. From increased food safety to higher profit margins, Stuart explains how AI in sorting is more than just a buzzword, and why food processors should pay attention.
Interviewer: Hi Stuart, thanks for joining. To begin with, can you explain to me ‘what is AI in relation to optical sorting?’
Stuart: When most people think about AI today, they think about generative AI, for example, ChatGPT. However, in optical sorting, we’re using a very specific and different area of AI, or machine learning, that focuses on image analysis. We’re looking at machine learning, deep learning and convolutional neural networks. The algorithm inside the sorter is using AI to analyse the images and create the best possible sort perfromance, and a level of quality that was not achievable before this technology.
Interviewer: My second question is ‘why?’ For example, If I’m a processor, and my machines are already working, and my manual workforce is doing the task, why should I be interested in AI and digitalization?
Stuart: Two answers to that. The first is that by using this technology, we can make you more productive and we can make you more money. Every organization has a challenging business environment. This includes, raw material prices, energy prices and supply chain resiliency issues. At times like this, it’s even more important to make your existing systems as efficient and productive as possible. Here’s where AI comes in:
Our AI sorter is achieving a +50% yield increase. This means customers will be making more good product out of the same amount of incoming product, enhancing both sustainability and profit.
Touching onto digitalization and why that’s important. Another way to improve profit is to maximize uptime. Some of the ways we can maximize uptime is with our connectivity solutions, for example, SORTEX Monitoring System, which allows you to connect to the machine and monitor the performance in real time. What gets monitored gets managed. By increasing transparency, processors access reduced energy, increased uptime, increased yield, and increased quality.
Interviewer: So, if I have this right, AI and digitalization is used to increase yield, uptime and profit? Do you have examples of a commodity where AI shows improvement in the yield?
Stuart: The first application we’re going to work with our new SORTEX AI700 is gluten-free oats. This means we must remove wheat, rye, and barley from the batch. While a traditional sorter does a good job with this application, an AI sorter does a much better job.
The AI will look at the grains, all individually, the barley, wheat, rye and oats, and it will look at ‘what makes up a barley’ it will look for all the commonalities, and it may come up with 2000, a long list of quality parameters that a human or existing technologies could never detect. This gives the machine a significantly improved ability to detect non-oats, and keep the oats gluten-free.
Interviewer: In the same way that generative AI can reduce the efforts of certain tasks, can AI in optical sorting reduce human effort and error?
Stuart: When you’re setting up and retraining a machine, apart from running the right product through the machine, everything else is automated, which means it is consistent, which gives you the best chance of hitting that 100% performance level. It’s true that these machines will require less human operation. It’s going down a similar journey to the automotive industry, from increasing driver assistance to semi-autonomous and eventually fully autonomous operation. It’s exciting. We’re seeing the advent of AI, which is changing optical sorting.
For more information about our latest SORTEX AI700 machine, please visit here.
