Coupling hyperspectral imaging with machine learning algorithms for detecting polyethylene (PE) and polyamide (PA) in soils

Huan Chen; Taesung Shin; Bosoon Park; Kyoung Ro; Changyoon Jeong; Hwang-Ju Jeon; Pei-Lin Tan

doi:10.1016/j.jhazmat.2024.134346

Coupling hyperspectral imaging with machine learning algorithms for detecting polyethylene (PE) and polyamide (PA) in soils

J Hazard Mater. 2024 Apr 18:471:134346. doi: 10.1016/j.jhazmat.2024.134346. Online ahead of print.

Authors

Huan Chen¹, Taesung Shin², Bosoon Park³, Kyoung Ro⁴, Changyoon Jeong⁵, Hwang-Ju Jeon⁵, Pei-Lin Tan⁶

Affiliations

¹ Department of Environmental Engineering and Earth Sciences, Clemson University, Clemson, SC 29634, USA; Biogeochemistry & Environmental Quality Research Group, Clemson University, Georgetown, SC 29442, USA.
² USDA Agricultural Research Service, US National Poultry Research Center, Athens, GA 30605, USA.
³ USDA Agricultural Research Service, US National Poultry Research Center, Athens, GA 30605, USA. Electronic address: bosoon.park@usda.gov.
⁴ USDA Agricultural Research Service, Coastal Plains Soil, Water & Plant Research Center, Florence, SC 29501, USA.
⁵ Red River Research Station, Louisiana State University Agricultural Center, Bossier City, LA 71112, USA.
⁶ Biogeochemistry & Environmental Quality Research Group, Clemson University, Georgetown, SC 29442, USA.

PMID: 38653139
DOI: 10.1016/j.jhazmat.2024.134346

Abstract

Soil, particularly in agricultural regions, has been recognized as one of the significant reservoirs for the emerging contaminant of MPs. Therefore, developing a rapid and efficient method is critical for their identification in soil. Here, we coupled HSI systems [i.e., VNIR (400-1000 nm), InGaAs (800-1600 nm), and MCT (1000-2500 nm)] with machine learning algorithms to distinguish soils spiked with white PE and PA (average size of 50 and 300 µm, respectively). The soil-normalized SWIR spectra unveiled significant spectral differences not only between control soil and pure MPs (i.e., PE 100% and PA 100%) but also among five soil-MPs mixtures (i.e., PE 1.6%, PE 6.9%, PA 5.0%, and PA 11.3%). This was primarily attributable to the 1st-3rd overtones and combination bands of C-H groups in MPs. Feature reductions visually demonstrated the separability of seven sample types by SWIR and the inseparability of five soil-MPs mixtures by VNIR. The detection models achieved higher accuracies using InGaAs (92-100%) and MCT (97-100%) compared to VNIR (44-87%), classifying 7 sample types. Our study indicated the feasibility of InGaAs and MCT HSI systems in detecting PE (as low as 1.6%) and PA (as low as 5.0%) in soil. SYNOPSIS: One of two SWIR HSI systems (i.e., InGaAs and MCT) with a sample imaging surface area of 3.6 mm² per grid cell was sufficient for detecting PE (as low as 1.6%) and PA (as low as 5.0%) in soils without the digestion and separation procedures.

Keywords: Hyperspectral imaging; Machine Learning; Microplastics; Polyamide; Polyethylene; Soil.