2025-12-09 –, Auditorium
In the Netherlands plastic waste is often gathered in a dedicated container. These are collected in truck, that are emptied at a transfer station. Here a visual inspection is done to identify stuff that does not belong in this waste. We have researched the automation of this inspection using cameras and vision foundation models. All items in image of the big pile are detected, segmented and classified whether they belong to this waste stream.
A lot of households nowadays have a dedicated container for plastic waste. It is not always fully clear for the public what belongs to this waste stream. So items not belonging in this container end up in there. After collecting the waste from the containers, the trucks are unloaded at the transfer station. Here there is a visual inspection of the truck load for items not belonging to this waste stream.
At the Nationaal Testcentrum Circular Plastics (NTCP) there was a project to find out if we could automate this process to have an objective method to inspect the waste. We decided to use cameras in a top view of the pile of items, as the inspection is done.
Per item, we first need to detect it in the image of the pile and segment the pixels of it. We build a pipeline using vision foundation models.
Once we have a separate image of each item we need to decide whether it is valid waste or an unwanted item. After spending a few days at the transfer station we found that the class of unwanted items had a too broad variety to build a classification network.
So we settled for using anomaly detection. We trained several models specialized for this detection. In the en using foundation models for visual together with textual features gave a good distribution of valid waste where we could apply out of distribution detection.
The results showed a good detection in the images, given the variety in both valid waste and unwanted items.
Tom Koopen is a highly experienced professional in computer vision and deep learning. He holds a Master’s Degree in Applied Physics from the University of Twente, specializing in optical measurement systems.
With over 25 years of experience in computer vision, Tom has worked with various companies in the Netherlands. Since 2013 he is an entrepeneur at “de tijdelijke expert”, assisting customers with the application of computer vision technology, focusing on measurements, identification, and sorting of products. Recently he founded “textilemining.eco” to build innovative machines for textile recycling.
He has designed lighting systems, selected and optimized cameras, and written software for thousands of hours. Some of his notable projects include inspecting plastic crates for contamination, improving the sorting of plastics, metals and flower bulbs. He developed a 3D scanner to recognize roof tiles for Luijtgaarden B.V. and measured colors during high-speed printing processes at QI Press Controls. Oh, and don’t forget the beer bottle inspection with 10 per second about 20 years ago.