bvse – AI as a Catalyst for the Materialkreislauf

Artificial Intelligence (AI) and Deep Learning are revolutionizing the recycling industry. If scaffolding is no longer efficient, finer sorting can take place. If you apply completely new applications. Dr. Volker Rehrmann, EVP and Manager of TOMRA Recycling, on the possibilities of AI for our industry. A TOMRA job article.

Dr. Volker Rehrmann, EVP and Leiter of TOMRA Recycling Bildquelle: TOMRAWe all stripes after a real Kreislauf economy. While this sets itself, the back-won material is a homogeneous high-quality and so much material that may well be in the Kreislauf is geführt. It is with the current paints not possible. A downcycling of the highest possible results, must be sorted.

The goal is: AI power is possible. Even more: AI would be a catalyst for the material cycle race. Make sure that the technology is expanded with the sorting and classification of recycling material. Materials are available and slow, so the strong demand for more recycling is contained in here.

Before we do the source acquisition of KI with one of our Frage women, the work is one of the following missverständnis broadcasts: KI is not the new fashionwort, but it is better to be in the Mittelpunkt-unserer Branche. Our Research and Development teams go to Jahren KI-gestützte sortierlösungen. The first TOMRA machines have been produced for 30 years in the Netherlands, and they were ready to be sorted, but the material was sorted into containers. And that separation, which affects a human being, is the most urgent definition of AI.

If we start in the year 2024, we can make our new discoveries in the research of Deep Learnings. There is a trade association for machine learning, which first by the Weiterentwicklung of the Rechenleistung is in the costs of an industry-funded hat for a year.

Who else has offered Deep Learning to resource recovery, who does he want to know?

1. Deep Learning offers more flexibility. If you are on a steep hike, a sorting system is used so that you can reach a new market place. Hardware components or other advanced machines can use modern Deep Learning technologies with regular software updates so that they can be trained by our experts. This allows us to respond faster to customer needs.

The sorting itself can be done easily. Original sorting systems have a number of brands that have a precise and effective operation, and also the trends of single and multi-shell PET shells. Combine with the best system, that is the best system with the following VIS sensors, with Deep Learning technology, with AUTOSORT™ with GAINnext™ from the fall, reach with the most advanced sorting granularity. You can sort with material art and color and now new Deep Learning of Form, Size, Measurements or other Details. So if you use an uncomplicated suction system, you can compare the sorting of the use of unused products with the use of unused options for PET, PP and HDPE. It is a Meilenstein for our industry, where GAINnext™ has become a hereditary purity grade of 95 Prozent for the Lebensmittelsicherheit in Europe.

3. Deep Learning is automated by analyses that we can perform. The work of Deep Learning lies in the object recognition with the help of the complete camera.

In other words: GAINnext™ sees what human eyes can see. As sorting is automated, manual processing operations must be scaled up to ensure that mixing and recycling of materials is fast and efficient.

​4. You can never make a big step towards process optimization with data help. AI-driven sorting systems generate general mixes and data about the material sisters, sorting efficiency and supply performance. While analyzing that data, you can recognize and act on optimal optimization methods. If the sorting system is more likely to be switched off. If you start working on the basis of Deep Learning chamber and learning processes in the sorting process, a process and material flow is executed in the Blik. So if you want to improve and thus ensure the quality of the sorting of the flow and the material consumption, the form scripts for the recycling of raw materials are one-off.

AUTOSORT™ with GAINnext™ Bildquelle: TOMRAAUTOSORT™ with GAINnext™

Our industry is also experiencing an exciting turning point. Since the deployment of AI is the cross-sectoral economy of a time point, it is urgently needed to – Legislative regulations are being issued and the customer demand for technologically advanced solutions is rising. Convergence is one of the times, the new market with higher production products and an industry that is further expanding. Go and see TOMRA, part of the AI ​​Revolution is ready!

What is Deep Learning?

Deep Learning is a part of AI. It is an imitation of art and well-being, where the human information is represented. You act with a special technique within machine learning, while the neural networks are trained, so that you can best recognize, save and that später on new data. TOMRA’s AI experts talk about millions of images as training material in the network, the best optical brand of material types that are specific bottle locks or packaging forms that are not separated. So Deep Learning can release a complex sorting, if it is the best combination sensors.
​
Source and further information: www.tomra.com/de-de