Estimation of pigment concentration in LDPE via in-line hyperspectral imaging and machine learning

  • Georgiana Amariei
  • Anne Sofie Schaarup-Kjær
  • Pernille Klarskov
  • Martin Lahn Henriksen
  • Mogens Hinge  
Plastic and Polymer Engineering, Department of Biological and Chemical Engineering, Aarhus University, Aabogade 40, DK-8200 Aarhus N., Denmark
[email protected]
 https://orcid.org/0000-0002-5412-6325
 Search for papers by this author
Plastic and Polymer Engineering, Department of Biological and Chemical Engineering, Aarhus University, Aabogade 40, DK-8200 Aarhus N., Denmark and American AVK Company, Quality Department, 2155 Meridian Blvd, Minden, NV 89423, USA
[email protected]
 https://orcid.org/0000-0003-3458-166X
 Search for papers by this author
Terahertz Photonics, Department of Electrical and Computer Engineering, Aarhus University, Finlandsgade 22, DK-8200, Aarhus N, Denmark
[email protected]
 https://orcid.org/0000-0002-9422-7841
 Search for papers by this author
Plastic and Polymer Engineering, Department of Biological and Chemical Engineering, Aarhus University, Aabogade 40, DK-8200 Aarhus N., Denmark
[email protected]
 https://orcid.org/0000-0002-5115-6166
 Search for papers by this author
 Corresponding Author
Plastic and Polymer Engineering, Department of Biological and Chemical Engineering, Aarhus University, Aabogade 40, DK-8200 Aarhus N., Denmark
[email protected]
 https://orcid.org/0000-0002-8787-5314
 Search for papers by this author

Due to the increasing amount of plastic waste and high-quality demands on recycled plastic interest for in-line composition estimation in plastics has grown the last few years. This study investigates pigment blue 15 : 3 with varying concentrations in LDPE. Samples are investigated with two industrial hyperspectral imaging systems where one has the hyperspectral range from 450 nm to 1050 nm and the other from 950 nm to 1750 nm. A model based on peak ratios of selected bands and model based on a principal component analysis have been tested. The models only predict pigment concentrations between 40.0 wt% and 1.7 × 10–3 wt% if both spectral ranges are combined. Unknown samples containing pigment concentration ranging from 20 wt% to 0.31 wt% were predicted and correlated to the actual pigment concentrations (R2 = 0.977) and the PC-based model outperforms the peak ratio model. The studied approach can be a part of the solution to the plastic challenge and can be transferred to other applications where concentration determination is key.

Keywords: hyperspectral imaging, pigment blue 15:3, pigment concentration, in-line concentration estimation, machine learning

Metrics

Downloads:

443

Abstract Views:

2,464