A methodology for the fast identification and monitoring of microplastics in environmental samples using random decision forest classifiers

Hufnagl, Benedikt; Steiner, Dieter; Renner, Elisabeth; Löder, Martin G. J.; Laforsch, Christian; Lohninger, Hans

doi:10.1039/c9ay00252a

Cited by 111 publications

(101 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Machine learning was also used to classify microplastics in spectral maps. Hufnagl et al 176 proposed a random forest approach to detect microplastics in mIR images as shown in Fig. 11.…”

Section: Machine Learning Applicationsmentioning

confidence: 99%

“…Hufnagl et al 176 claim that the algorithm (a) (b) (c) (d) Figure 11. Microplastic classification in mIR images using a machine learning approach introduced by Hufnagl et al 176 The images from left to the right represent a zoomed-in and panned view of a sample: (a) optical image of the microplastic sample, (b) mIR image, (c) pixelwise classification output of the machine learning model for the different polymer types (including PE, PP, PMMA, PAN, PS, and nonpolymer particles), and (d) overlay of the classification result with the optical image of the sample. Reproduced from Hufnagl et al 176 with permission from The Royal Society of Chemistry.…”

Section: Machine Learning Applicationsmentioning

confidence: 99%

See 1 more Smart Citation

Critical Review of Processing and Classification Techniques for Images and Spectra in Microplastic Research

et al. 2020

View full text Add to dashboard Cite

“…Machine learning was also used to classify microplastics in spectral maps. Hufnagl et al 176 proposed a random forest approach to detect microplastics in mIR images as shown in Fig. 11.…”

Section: Machine Learning Applicationsmentioning

confidence: 99%

Section: Machine Learning Applicationsmentioning

confidence: 99%

Critical Review of Processing and Classification Techniques for Images and Spectra in Microplastic Research

et al. 2020

View full text Add to dashboard Cite

“…This identification is quick and allows the user to judge the interpretations made by the machine. The KNN method applied to IR spectra could also be applied to data obtained using automated methods such as focal plane array (FPA) detector combined with BaF2 window or, with another learning database, to data from Raman spectrometry (Cabernard et al, 2018;Hufnagl et al, 2019;Primpke et al, 2019). Raman spectrometry is indeed often used for the analysis of microplastics and is likely to provide additional information (Frère et al, 2017;Hahn et al, 1997;Hiejima et al, 2018;Zhao et al, 2017).…”

Section: Accepted Manuscriptmentioning

confidence: 99%

“…In 2018, an automated identification method based on hierarchical cluster analysis was published (Primpke et al, 2018). More recently a new method based on Random Decision Forest was proposed (Hufnagl et al, 2019).…”

Section: Introductionmentioning

confidence: 99%

“…Recently, a machine learning process has been successfully used on the determination of the chemical nature of microplastics (Hufnagl et al, 2019). Among these methods, one of them, called K-Nearest Neighbors (KNN), is known for its simplicity and efficiency in automated spectrum recognition (Ripley, 1996;Venables et al, 2002;Wehrens, 2011).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A machine learning algorithm for high throughput identification of FTIR spectra: Application on microplastics collected in the Mediterranean Sea

et al. 2019

View full text Add to dashboard Cite

The development of methods to automatically determine the chemical nature of microplastics by FTIR-ATR spectra is an important challenge. A machine learning method, named k-nearest neighbors classification, has been applied on spectra of microplastics collected during Tara Expedition in the Mediterranean Sea (2014). To realize these tests, a learning database composed of 969 microplastic spectra has been created. Results show that the machine learning process is very efficient to identify spectra of classical polymers such as poly(ethylene), but also that the learning database must be enhanced with less common microplastic spectra. Finally, this method has been applied on more than 4000 spectra of unidentified microplastics. The verification protocol showed less than 10% difference in the results between the proposed automated method and a human expertise, 75% of which can be very easily corrected. Highlights ► A machine learning algorithm was developed to determine the chemical nature of microplastics. ► This method allows a fast and reliable automated identification even when several thousand of FTIR spectra have to be studied.► This method is the first part of a software dedicated to the study of microplastics: POSEIDON.

show abstract

Tackling the Challenge of Extracting Microplastics from Soils: A Protocol to Purify Soil Samples for Spectroscopic Analysis

Möller

Heisel

Satzger

et al. 2021

Enviro Toxic and Chemistry

Self Cite

View full text Add to dashboard Cite

Microplastic pollution in soils is an emerging topic in the scientific community, with researchers striving to determine the occurrence and the impact of microplastics on soil health, ecology, and functionality. However, information on the microplastic contamination of soils is limited because of a lack of suitable analytical methods. Because micro-Fouriertransform infrared spectroscopy (µ-FTIR), next to Raman spectroscopy, is one of the few methods that allows the determination of the number, polymer type, shape, and size of microplastic particles, the present study addresses the challenge of purifying soil samples sufficiently to allow a subsequent µ-FTIR analysis. A combination of freeze-drying, sieving, density separation, and a sequential enzymatic-oxidative digestion protocol enables removal of the mineral mass (>99.9% dry wt) and an average reduction of 77% dry weight of the remaining organic fraction. In addition to visual integrity, attenuated total reflectance FTIR, gel permeation chromatography, and differential scanning calorimetry showed that polyamide, polyethylene, polyethylene terephthalate, and polyvinyl chloride in the size range of 100 to 400 µm were not affected by the approach. However, biodegradable polylactic acid showed visible signs of degradation and reduced molecular weight distribution after protease treatment. Nevertheless, the presented purification protocol is a reliable and robust method to purify relatively large soil samples of approximately 250 g dry weight for spectroscopic analysis in microplastic research and has been shown to recover various microplastic fibers and fragments down to a size of 10 µm from natural soil samples.

show abstract

A methodology for the fast identification and monitoring of microplastics in environmental samples using random decision forest classifiers

Abstract: A new yet little understood threat to our ecosystems is microplastics.

Cited by 111 publications

References 26 publications

Critical Review of Processing and Classification Techniques for Images and Spectra in Microplastic Research

Critical Review of Processing and Classification Techniques for Images and Spectra in Microplastic Research

A machine learning algorithm for high throughput identification of FTIR spectra: Application on microplastics collected in the Mediterranean Sea

Tackling the Challenge of Extracting Microplastics from Soils: A Protocol to Purify Soil Samples for Spectroscopic Analysis

Contact Info

Product

Resources

About