Detecting Contaminants in Post-Consumer Plastic Packaging Waste by a Nir Hyperspectral Imaging-Based Cascade Detection Approach

Bonifazi, Giuseppe; Gasbarrone, Riccardo; Serranti, Silvia

doi:10.31025/2611-4135/2021.14086

Cited by 8 publications

(9 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The effectiveness of HSI for solid waste classification has been verified in prior research. 17,18 However, to the best of our knowledge, no study has investigated the potential HSI combined with ensemble ML algorithms to quantitatively detect the carbon groups and contents of solid waste. This is the first study to build chemical maps to show the distribution of these constituents in a straightforward and simple manner, which confirms the acceptable accuracy.…”

Section: ■ Results and Discussionmentioning

confidence: 99%

“…16,17 For instance, recent studies classified construction wastes (bricks, concretes, woods, plastics, and rubbers) for sorting and recycling 17 and identified impurities (paper and carboard) from plastic packaging waste streams. 18 However, HSI-aided classification technology is not yet fully developed. The spectral features of a huge variety of solid waste samples in the NIR range are not well characterized.…”

Section: ■ Introductionmentioning

confidence: 99%

“…It has proven to be a powerful technique applied extensively in the agricultural, pharmaceutical, and food industries. Currently, several studies on NIR HSI for solid waste separation and recycling have attracted intensive attention. , For instance, recent studies classified construction wastes (bricks, concretes, woods, plastics, and rubbers) for sorting and recycling and identified impurities (paper and carboard) from plastic packaging waste streams . However, HSI-aided classification technology is not yet fully developed.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Machine Learning and Hyperspectral Imaging-Aided Forecast for the Share of Biogenic and Fossil Carbon in Solid Waste

Lan

et al. 2023

ACS Sustainable Chem. Eng.

View full text Add to dashboard Cite

Traditional methods for analyzing the biogenic and fossil carbon shares in solid waste are time-consuming and labor-intensive. A novel approach was developed to directly classify the carbon group and predict carbon content using the hyperspectral imaging (HSI) spectra of solid waste in conjunction with state-of-the-art tree-based machine learning models, including random forest (RF), extreme gradient boost, and light gradient boost machine (LGBM). All of the classifiers and regressors were able to achieve an accuracy above 0.95 and an R 2 of 0.96 in the test set, respectively. In addition, two model interpretation approaches, the Shapley additive explanation and model explainer, were applied. The results showed that the predictions of the developed models were based on a reasonable understanding of the overtone and shake of the functional groups (C–H, N–H, and O–H). Furthermore, the developed models were validated by an external test set, which did not overlap with the data used for model construction. The RF and LGBM showed robust performance with a 0.790 accuracy for carbon group classification and a 0.806 R 2 for carbon content prediction. Overall, the optimal models provided a rapid method for characterizing the biogenic carbon share in solid waste based on raw HSI spectra without preprocessing.

show abstract

Section: ■ Results and Discussionmentioning

confidence: 99%

Section: ■ Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Machine Learning and Hyperspectral Imaging-Aided Forecast for the Share of Biogenic and Fossil Carbon in Solid Waste

Lan

et al. 2023

ACS Sustainable Chem. Eng.

View full text Add to dashboard Cite

show abstract

“…However, c and C cannot be routinely measured today. With the development of combined optical measurements of the composition and size of individual particles and machine learning techniques, also this approach considering the of mass fraction of particles of interest could be feasible in the future (Bonifazi et al, 2021;Kroell et al, 2021).…”

Section: Requirement Ii: Minimum Numbers Of Particles "Rare In Size" ...mentioning

confidence: 99%

A novel method to calculate the size of representative waste samples based on particles size

Beggio

Hennebert

2022

Detritus

View full text Add to dashboard Cite

A novel approach to determine the size of samples of granular wastes is proposed, forwarding the concept of the “number of particles”, as previously introduced by the authors. To be representative with a minimum error, it was demonstrated that at least 100 particles showing the presence of the measurand, shall be collected in the sample. Waste particles are usually characterized by size-concentration relationships. However, in waste sampling standards they are not explicitly considered when estimating the size of the sample. In this context, this paper extends this requirement to the number of particles “rare is size”, belonging to the less represented size fraction in the waste to be characterized. The number of particles is then transformed into a mass by a formulation that avoids using unrealistic assumptions on particles features. Results derived from the application of the two formulations on 5 different types of waste show that their equivalency relies on how similar are, the proportions of particles rare in concentration and rare in size in the batch to analyse. Here, preliminary knowledge on particles physical features and distribution of the measurand is key to derive coherent values for mass of samples. Finally, the need to perform on-site size reduction is discussed for cases where the application of both the conventional and novel approaches could have leaded to unpractical management of too large-sized waste samples.

show abstract

“…Within this study, the suitability of the existing sorting infrastructure for the processing of bioplastic waste was assessed. For this purpose, NIR sorting technology, which is a widely used plastic sorting technology material (Bonifazi et al, 2021;Helena Wedin et al, 2017;Zhu et al, 2019), was used to check the sortability of bioplastics.…”

Section: Introductionmentioning

confidence: 99%

Near-Infrared Identification and Sorting of Polylactic Acid

2022

View full text Add to dashboard Cite

Biobased plastics are often seen to be an environmentally friendly alternative to conventional plastics, with their share, though being less now, is gradually increasing. This necessitates that the waste management of these possibly eco-friendly materials is also at par with their growth. Near-infrared (NIR) sorting is an effective waste sorting technology and is already widely used for conventional plastics. Thus, it would be imperative to analyse whether this effective existing infrastructure could also be successfully used to sort bioplastic. In the present study, the lab-scale NIR sensor-based sorting system in Montanuniversität Leoben was used to analyse polylactic acid (PLA) in three sets of experiments. First, the spectra of 7 conventional plastics were compared to that of virgin PLA and it was found that PLA has a distinct spectrum and should ideally be detected from a mixed plastic fraction. Second, it was assessed whether different grades and thicknesses of virgin PLA samples produced different spectra and it was found that there is a slight difference in the intensities without any wavelength shift of the recognizable peaks. Lastly, the detection of 10 PLA product samples was tested using the NIR recipe of a virgin PLA. It was observed that the samples were successfully detected and blown out as PLA for all the conducted trials. Additionally, it was also seen that an appropriate backlight setting is important to be able to correctly sort the transparent PLA products in the used chute-type sorter.

show abstract

Detecting Contaminants in Post-Consumer Plastic Packaging Waste by a Nir Hyperspectral Imaging-Based Cascade Detection Approach

Cited by 8 publications

References 24 publications

Machine Learning and Hyperspectral Imaging-Aided Forecast for the Share of Biogenic and Fossil Carbon in Solid Waste

Machine Learning and Hyperspectral Imaging-Aided Forecast for the Share of Biogenic and Fossil Carbon in Solid Waste

A novel method to calculate the size of representative waste samples based on particles size

Near-Infrared Identification and Sorting of Polylactic Acid

Contact Info

Product

Resources

About