Can Recycling
Open Learning/Innovation
Client company:Tomra
Zoe Josef Despo Maria Spyrou
Kyllian Warmerdam
Cindy Huijskens
Nikola Kličková
Jordi Coll Moreno
Project description
The primary goal of this project is to enhance the existing can recycling processes by developing a smart AI-based disassembly system that can recognize the deposit value of cans, even if they are deformed. Current systems require cans to be in their original shape for barcode recognition, which limits efficiency. By enabling the recognition of cans in various deformed states, we aim to offer economic and environmental benefits, such as reduced CO2 emissions, improved storage efficiency, and greater recycling capacity.
This innovation is in line with the broader sustainability objectives of both Tomra and Fontys University, aiming to improve recycling practices and support a circular economy.
Context
The primary goal of this project is to enhance the existing can recycling processes by developing a smart AI-based disassembly system that can recognize the deposit value of cans, even if they are deformed. Current systems require cans to be in their original shape for barcode recognition, which limits efficiency. By enabling the recognition of cans in various deformed states, we aim to offer economic and environmental benefits, such as reduced CO2 emissions, improved storage efficiency, and greater recycling capacity.
This innovation is in line with the broader sustainability objectives of both Tomra and Fontys University, aiming to improve recycling practices and support a circular economy.
Results
The project utilized various advanced technologies to enhance can recycling processes through an AI-based disassembly system. We focused on depth vision to analyze can shapes, sound detection to identify structural integrity, and AI object recognition for classifying cans based on their condition. Throughout the sprints, we gathered a dataset of images and sounds from both intact and crushed cans, developed machine learning models, and designed a prototype machine for testing. Additionally, we integrated hardware components with AI models to create a functional prototype aimed at improving recycling efficiency and supporting sustainability goals.