SORTING | TECHNOLOGY


and Vision spectroscopy to separate them by colour, with above 99% purity for the clear PET fraction. Once sorted, PET bottles (around 70%) and aluminium cans (around 20%) are sent to their own recycling channels for a second life. The plant has capacity to process more than 14,000 tonnes of PET bottles and beverage cans every year. Like other optical sorting technology compa-


nies, Pellenc ST has been developing AI. Last year, it launched its first solutions equipped with AI for the cleaning of PE streams and the complex separation of paper/cardboard. By relying on its CNS Brain technology, Pellenc ST is now taking up a new challenge in the field of PET sorting. “We have developed a new artificial intelligence model capable of recognising transparent PET bottles covered with a sleeve or a large label, which are difficult to recognise using NIR/VIS alone. Our model already detects more than 85% of sleeved bottles and we expect to exceed the 90% mark in the near future,” said Kevin Alazet, AI manager for Pellenc ST. This solution also helps to reduce losses, which have fallen from 7% to less than 2%. CNS Brain is one of the AI options offered by Pellenc ST that does not require additional equipment and does not generate additional energy consumption or maintenance needs, the company says. For more difficult applications, Pellenc ST equips its machines with an additional camera and offers the AISort option to address new use cases. The best example, according to the company, is the separa- tion of food packaging and non-food packaging made from the same or similar material. Thanks to the AI module and multi-sensor fusion, the optical sorters supplied by Pellenc ST can now address this issue and obtain high-quality recycled products through extremely accurate sorting, it says.


Future directions What if sorting of multi-layer materials is too hard a nut to crack? This is an oversimplification of a review from a team of researchers at the University of Buffalo in the US in Industrial & Engineering Chemistry Research, which examines how emerg- ing technologies can make plastic recycling more efficient.


Using a process-systems-engineering frame-


work, the research considers a range of techniques, from chemical solvents that selectively dissolve particular polymers to AI-enhanced automated sorting systems capable of improving separation of complex waste streams. The authors conclude that solvent-based recycling represents a promising, economically viable path (particularly for plastics


www.plasticsrecyclingworld.com


traditionally hard to sort) though replacing fossil- derived plastics with bio-based alternatives remains uncertain for now. In May this year, a research team led by Vaisha-


li Maheshkar (also from University of Buffalo) and colleagues released a detailed evaluation of machine-learning methods for optical plastic sorting in MRFs. The study compiled a novel dataset exceeding 20,000 images – drawn from real-world MRF feeds, public collections, and online sources – to thoroughly test computer-vision models including two-stage detectors like Mask R-CNN and single-stage models like YOLO. The researchers’ goal was to assess both the


strengths and practical limitations of these visual systems under realistic operating conditions. The research found that although optical systems can identify plastics under controlled conditions, they struggle in real world environments because they overly depend on visual cues like colour, shape, labels, or even the conveyor belt itself, rather than intrinsic material properties. The study therefore recommends the integration of complementary sensing methods – such as spectroscopy – with computer vision to achieve accurate and reliable sorting of plastics in complex real-world waste streams.


CLICK ON THE LINKS FOR MORE INFORMATION: � www.tetrapak.com � https://recycleye.com � www.mssoptical.com � www.tomra.com � www.greyparrot.ai � https://prezero-international.com � www.wesort.ai � https://omraa.no/en � https://infinitum.no � www.nordic-recycling.com � www.pellencst.com � www.buffalo.edu � www.digitalwatermarks.eu


September 2025 | PLASTICS RECYCLING WORLD 35


Above: Pellenc ST Compact+ optical sorting unit


IMAGE: PELLENC ST


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  |  Page 45  |  Page 46