{"id":1078,"date":"2021-08-19T11:53:00","date_gmt":"2021-08-19T09:53:00","guid":{"rendered":"https:\/\/flikweertvision.nl\/?p=1078"},"modified":"2026-08-25T12:39:32","modified_gmt":"2026-08-25T10:39:32","slug":"flikweert-vision-deploys-artificial-intelligence-for-sorting","status":"publish","type":"post","link":"https:\/\/flikweertvision.blackdesk.dev\/en\/flikweert-vision-zet-kunstmatige-intelligentie-in-voor-sortering\/","title":{"rendered":"Flikweert Vision deploys artificial intelligence for sorting"},"content":{"rendered":"
\n
\n \"Flikweert\n
\n
\n
\n
\n \n News\n <\/span>\n

Flikweert Vision deploys artificial intelligence for sorting<\/h1>\n

Flikweert Vision deploys artificial intelligence for sorting<\/h1>\n
\n 19 August 2021<\/span>\n <\/div>\n <\/div>\n
\n \n Scroll<\/span>\n <\/i>\n <\/a>\n <\/div>\n <\/div>\n <\/div>\n <\/div>\n <\/figure>\n <\/section>\n
\n <\/i><\/a>\n <\/div>\n
<\/div>\n \n\n
\n
\n
\n
\n
\n QualityGrader<\/span>\n <\/div>\n

Flikweert Vision deploys artificial intelligence for sorting<\/h2>\n

Flikweert Vision deploys artificial intelligence for sorting<\/h2>\n

The application of artificial intelligence aims to make many manual operations in sorting seed potatoes unnecessary. By using cameras, Flikweert Vision Technologies, based in Ouwerkerk, Zeeland, is making it possible to largely automate the sorting of difficult-to-sort batches based on quality. The first version of the compact QualityGrader, which is 3.3 metres long and 1.2 metres wide, is currently being extensively tested at arable farm Quaak Potato in Scherpenisse. \u201cThe focus is on optimising quality assessment and separation between good and bad potatoes.\u2019<\/p>\n

Automation in ware potatoes<\/strong><\/p>\n

As the sorting season for seed potatoes came to an end, Flikweert chose to use the machine for sorting ware potatoes. That Quaak Potato has now decided to purchase three QualityGraders indicates that the entrepreneur is convinced of the added value. Three sorting tables have been replaced by the young Ouwerkerk-based company's QualityGraders. It is estimated that each machine sorts about 10 tonnes of potatoes per hour. \u201cThis enables us to process about one trailer per hour, which is the desired capacity,\u201d says Martijn Flikweert.<\/p>\n

Benefits of artificial intelligence in agriculture<\/strong><\/p>\n

The application of artificial intelligence in agriculture has been a quest. After graduation, Martijn Flikweert and his cousin Lars Flikweert decided to establish Flikweert Vision Technologies in 2019. Both have a background in arable farming and studied Agrotechnology & Management in Dronten and Biosystems Engineering in Wageningen, respectively. \u201cWe saw that artificial intelligence could be of great value in a sector where a lot of manual work is still done. After all, agriculture works with natural products, where seasonal changes and weather conditions affect the products. It is these variations in natural products that convinced us that artificial intelligence can add significant value.\u201d<\/p>\n

Take over reading<\/strong><\/p>\n

Although initially thinking of sorting Brussels sprouts, Lars and Martijn Flikweert shifted their focus to potatoes. To arrive at the optimal grading machine, based on the developed software, Stefan Flikweert, Martijn's brother, was involved in the project. \u201cStefan, who studied mechanical engineering, took on the development and design of the QualityGrader,\u201d says Martijn.\u201d<\/p>\n

\u201cOur starting point was to help a seed potato grower who is happy with his existing grader but wants to spend less time in the grading room. So what did the QualityGrader have to be able to do? We deliberately chose sorting based on quality. What we often see from current optical sorting machine providers is that these machines focus on accurate sizing. While this is certainly valuable, we believe that many growers mainly want to automate the reading work. This was the reason for developing a compact and affordable machine that takes over much of the reading work.\u201d<\/p>\n

Recognise scab, damage, growth cracks and mould defects<\/strong><\/p>\n

Initially, a prototype was developed with a second-hand roller conveyor and a row of shoot-out scoops behind it. \u201cBy looking critically at the prototype, we learned a lot, both software-wise and mechanically.\u201d With the improvements that emerged from the prototype, drawings were made for the first QualityGrader. As there were several growers in the area with lots of seed potatoes that had a lot of scab, the focus was mainly on detecting scab.<\/p>\n

The difficulty in detecting scab, according to Martijn Flikweert, also marked the limit of the optical grading machine's capabilities. \u201cIf we can detect scab, we can also detect damages, shape defects and growth cracks, we thought. This proved to be true, and because we now focus mainly on consumption potatoes, we can now accurately identify other defects besides scab.<\/p>\n

Generating data<\/strong><\/p>\n

The optical sorter is equipped with three industrial cameras that read the potatoes. \u201cThe length of the machine is needed to properly position the potato in front of the camera,\u201d says Dr. K. The algorithm becomes more accurate as more QualityGraders are installed. \u201cWe create a dataset. Examples of potatoes with different defects lead to an assessment form. This form acts as the algorithm used to assess each individual potato.\u201d Each potato is photographed five times from different angles. The software then determines which defects are present or missing. \u201cThis system is still under development, but that mainly depends on generating and processing more data. This will result in a more consistent, accurate and robust assessment.\u201d Plastic scoops at the back of the grading machine ultimately separate bad potatoes from good potatoes.<\/p>\n

The user ultimately determines the desired precision of the selection. \u201cThis can be tailored to each lot. For example, based on export targets or lot classes. Different defects can be set separately. If a user also wants to detect small defects, the damage threshold should be lowered.\u201d<\/p>\n

Advantages of the QualityGrader<\/strong><\/p>\n

Interest in the QualityGrader is now so high that several growers have proceeded to purchase it after demonstrations. \u201cWe hope to install six or maybe even more QualityGraders this year,\u201d says Martijn Flikweert's ambition. The main advantage, of course, lies in the saving of labour in reading. \u201cAt the moment, there are people who don't even pick out every bad potato. In practice, people's attention span decreases as the day progresses. This leads to quality differences and varying processing capacity. What percentage of potatoes are currently sorted correctly by the QualityGrader is difficult to say. This currently depends on the batch and the desired settings. This is also the reason why our full focus in the coming period is on improving the accuracy of decisions and minimising the impact of lot differences on performance.\u201d<\/p>\n

\u201cFor our target group, the average seed potato grower and potato processor who is satisfied with the current size grading but wants to spend less time in the grading room, we want to offer an affordable machine.\u201d Martijn Flikweert emphasises that Flikweert Vision Technologies\u201c efforts remain entirely focused on optical sorting technology for the time being. \u201dWe want to develop and refine a specialised solution for reading automation. If several growers are satisfied with the operation of a QualityGrader, we hope to scale up production further. Ultimately, the complexity lies in the software. The initial separation has to be as optimal as possible. Although we have already made significant progress in this, we aim for even more robustness and reliability.\"<\/p>\n

Original article: Arable.co.uk<\/p>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n <\/section>","protected":false},"excerpt":{"rendered":"","protected":false},"author":2,"featured_media":1079,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1],"tags":[],"class_list":["post-1078","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-nieuws"],"acf":[],"_links":{"self":[{"href":"https:\/\/flikweertvision.blackdesk.dev\/en\/wp-json\/wp\/v2\/posts\/1078","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/flikweertvision.blackdesk.dev\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/flikweertvision.blackdesk.dev\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/flikweertvision.blackdesk.dev\/en\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/flikweertvision.blackdesk.dev\/en\/wp-json\/wp\/v2\/comments?post=1078"}],"version-history":[{"count":4,"href":"https:\/\/flikweertvision.blackdesk.dev\/en\/wp-json\/wp\/v2\/posts\/1078\/revisions"}],"predecessor-version":[{"id":2258,"href":"https:\/\/flikweertvision.blackdesk.dev\/en\/wp-json\/wp\/v2\/posts\/1078\/revisions\/2258"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/flikweertvision.blackdesk.dev\/en\/wp-json\/wp\/v2\/media\/1079"}],"wp:attachment":[{"href":"https:\/\/flikweertvision.blackdesk.dev\/en\/wp-json\/wp\/v2\/media?parent=1078"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/flikweertvision.blackdesk.dev\/en\/wp-json\/wp\/v2\/categories?post=1078"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/flikweertvision.blackdesk.dev\/en\/wp-json\/wp\/v2\/tags?post=1078"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}