Hyperspectral imaging based techniques in fluff characterization

Light fractions produced after vehicles dismantling are conventionally defined as "fluff" or Automotive Shredder Residue (ASR). They represent about the 25% of the weight of a car and are usually constituted by materials characterized by intrinsic low specific gravity (i.e. plastics, rubber, synthetic foams, etc.). Fluff is usually polluted by metal contaminants (i.e. copper, aluminum, brass, iron, etc.), that strongly affect, especially in the final fractions, the possibility to utilize such material as fuel in co-combustion process, reducing the waste disposal and increasing at the same time energy production. In this paper, innovative selection-control architectures, based on hyperspectral imaging, in the visible- near infrared (VIS-NIR) field, have been investigated. In order to define suitable inspection strategies for the recognition and separation between useful (fluff) and polluting (metals) materials, samples of light and heavy plastics and metals have been collected in a recycling plant. Reflectance spectra have been acquired in the VIS-NIR field (400-1000 nm). Results showed as the different materials are characterized by different spectral signatures and that recognition of plastics and metals can be obtained adopting a wavelength ratio in the NIR (700-1000 nm) field.