Plastic solid waste identification system based on near infrared spectroscopy in combination with support vector machine

In this paper, identification system of plastic solid waste (PSW) based on near-infrared (NIR) reflectance spectroscopy in combination with Support Vector Machine (SVM) was presented. A device applied to obtain NIR spectra of plastics in the detection platform was developed. After pre-processing (normalized, 1st derivative and smooth), the repeatability of spectral absorption features was improved, which would assist the identification. A “principal component analysis (PCA)SVM” identification method was proposed to identify polypropylene (PP), polystyrene (PS), polyethylene (PE), poly(methyl methacrylate) (PMMA), acrylonitrile butadiene styrene (ABS) and polyethylene terephthalate (PET) among plastics, and its identification accuracy can reach 97.5%. The type of samples could clearly be identified and the shape of samples could also be roughly discerned. It is clearly shown that this system can achieve good identification results while reducing costs considerably, which has great potential in industrial recycling. Keywords: Plastic identification, Near-infrared spectroscopy, Principal component analysis, Support vector machine