With robots, workers can be replaced in order to lower labour costs, but that’s not the most important aspect of robotisation, according to Richard van de Linde of robot builder FTNON. The digitisation of the sector continues, and the importance of quality and traceability are only increasing, he said yesterday during the meeting of the Agro Businessclub Westland. That opportunities to that end are also increasing, became apparent from the explanation of Herbert ten Have of Delft Robotics, who is developing artificial intelligence for the fresh produce sector, among other sectors.
The meeting was held at Rabobank Westland.
Robotisation
The wildest stories are circulating about it. Two computers creating such an advanced language the plugs have to be pulled. DCs without human workers, and a future in which we are ruled by machines. Yesterday during the meeting of the Agro Businessclub Westland, the robotisation of the fresh produce sector was discussed.
Richard van de Linde started Lacquey while he was a student, known for the bell pepper sorting machine with the accurate grabbers that’s in operation at Haluco and Combilo. In 2015 the company was taken over, and as part of FTNON, the lettuce and cabbage sector is now on the agenda. According to Richard’s calculations, their recently developed machine is a closed business case, yet that isn’t his most important argument to take robotisation in the sector seriously.
IT in fresh
He indicates the major players in technology, such as Amazon and Google, that are rushing into the fresh sector. “These kinds of major players will decide the rules and have high requirements for digitisation and quality,” he predicts. “The information about where your product’s from and how it was processed is becoming increasingly important.” According to him, that’s where machines can actually prove their added value.
Ronald Grootscholten counts five new names on the club’s member list.
Richard van der Linde, FTNON
Why lettuce?
That FTNON is now focused on the lettuce sector is because of the enormous amount of product globally, the major players in it and the opportunities for adding value: the price of a head of lettuce is a fraction of the price of a meal salad. Moreover, it’s a difficult sector: no two products are the same, and the hygiene and cleaning requirements placed on the machines are high. Because of that, the number of competitors is limited. The company has now developed a machine that is five times faster and four per cent more accurate at removing the heart from the lettuce. To that end, five cameras make an accurate 3D model of each head. The machine is in operation at large lettuce processors in the Netherlands and the US by now, and Richard expects to sell more in coming time. After that, the company will focus on removing the core from cabbages. “Work that’s difficult to do manually, and a product that’s processed in large numbers as well,” he explains the choice.
The difference in value of a head of lettuce compared to a meal salad.
Learning
Conducting operations is one part, but how can a machine handle all the different products and information? That is the field of Herbert ten Have, from Delft Robotics. The company is specialised in developing systems that can handle varied products based on 3D vision: different types of boxes, for example, but also for the fresh sector.
“It isn’t possible to teach a robot all possible bell peppers in the world,” Herbert says. “Or all types of phalaenopsis. The product is too diverse for that.” For that, deep learning is needed. The company develops algorithms that are getting better and better thanks to artificial intelligence. The result is software that can, for example, count phalaenopsis buds and flowers, and it can do so better than humans. “If you can make an image of something, we can often translate that into information,” Herbert explains.
But if that’s possible, where’s the company’s harvesting robot? Herbert explains that what they do is the first stage of the process. “When you have this information, you need the right grabber to execute the movements. And that’s when integration comes into play: the machine has to function in a humid, warm environment.” And besides, all of that has to result in a closing business case. “We create the transition of image to information by order.”
Drinks after the meeting. Pictured are Richard van der Sande from Certhon, Ted Vollebregt from Klasmann Deilmann and Gert-Jan van Geest.
Gerard van den Bos, Carcoboss / Van Daalen, Ronald Arkestijn from Zyon Group and Ronald Hirschmann from ABN AMRO.
Cees Uitbeijerse from BDO Accountants and Peter Quist from Westland Partners.
Jan-Willen de Vries from Wageningen UR, Herbert ten Have from Delft Robotics and Henriëtte Hoogstad-Bakker, ING Grootzakelijk.
Ronald Grootscholten from FloraPartners, Martijn van Andel from Jem-ID and Bert van Ruijven from Arcadia Chrysanten.
Kees van Geest, Robin Berendse from Yeald.
Herwi Rijsdijk, ABC Westland, Edwin Vanlaerhoven, Best Fresh Group and Piet Zwinkels, Zwinbrothers.
Marcel Kortekaas, Looye Kwekers, John Duynisveld from ERP.
Menno Reijgersberg from ERP and Ger van Burik, Holland Fresh Group.