Hyperspectral imaging applied to quality control of end-of-life plastics waste

The production of end-of-life (EOL) polymer waste has dramatically increased over the last 60 years. The most important polymers used in industry are: low-density polyethylene (LDPE), polyamide (PA), high-density polyethylene (HDPE), polyethylene terephthalate (PET), polypropylene (PP), polyoxymethylene (POM), polystyrene (PS), polyvinyl chloride (PVC). This large number of produced plastic materials, characterized by different properties, led to an uncontrolled growth in production generating large amounts of plastic wastes, thus leading to a not any more sustainable plastic use (Bonifazi et al., 2009). Polymeric waste should be thus considered, more and more, as a resource for new products manufacturing, applying recycling strategies (Vilaplana and Karlsson, 2008)]. Several efforts can be made to achieve this qualitative step in plastic recycling, and consequently to introduce high quality recycled products into the market. Recycling processes require high degrees of separation and sorting in order to obtain a recycled polymer with a performance and quality not so different from those of the virgin polymers. Furthermore, an accurate quality certification of products, directly carried out during recycling processes, can give them a higher economic value (Luciani et al., 2013). In this perspective, the realization of a suitable sensor technology able to identify different types of plastics in a single one step can play a key role for the development of quality control strategies specific for this sector. HyperSpectral Imaging (HSI) is a technological platform that integrates conventional imaging and spectroscopy (Gowen et al., 2007)) and can be profitably utilized to set up quality control actions to apply in the plastic recycling sector. In this work, an innovative strategy allowing to recognize LDPE, HDPE and other polymers in a plastic waste flow stream was explored. A full classification of post-consumer plastics was carried out and a hierarchical classification method, based on Hyperspectral Imaging (HSI) in the SWIR (1000-2500 nm) range, was developed, set up and implemented. The proposed HSI approach has many advantages, being fast, non-destructive and accurate, without any need to perform specific sample collection and/or preparation.