Norwegian AI Directory
Description:
PlasticSort was born directly from the need of an end-user: RETUC, automotive thermoplastic recycler, is performing the task of plastic separation by hand, which kept proving to be both inefficient and prone to errors. Through their close relationship with Andaltec they learnt that no current solutions for this issue existed in the market, and so, they looked for other companies that could assist in providing a solution, hence forming the current Plasticsort consortium: ISR was contacted for their experience in sensor vision and algorithmic control of machines, NEO for their tailor-made solutions in hyperspectral censoring, and POLYCOM as a company that could verify this technology in real life by manufacturing new automotive components from recycled plastics. PlasticSort will solve this problem by providing the automotive plastic component manufacturers and recyclers with an innovative sorting machine for high quality thermoplastics which efficiently provides extremely pure sorted plastic appropriate for recycling.
PlasticSort is a multisensorial sorting machine which operates by processing information from NIR hyperspectral camera, visible light camera and inductive sensors using self-learning multi-layer perception neural network analysis. The basis of its material separation concept can be described referring to its three main aspects: preparation, classification and sorting.
NEO will be in charge of the hyperspectral sensor for the project.
Project leader: Trond Løke
Started: 2016
Ends: 2016
Category: Næringsliv
Sector: Næringsliv
Budget: 39999
Institution: NORSK ELEKTRO OPTIKK AS
Address: