The Identification of Spherical Engineered Microplastics and Microalgae by Micro-hyperspectral Imaging.

Area of Science:

• Environmental Science
• Analytical Chemistry
• Marine Biology

Background:

• Microplastic pollution is a growing environmental concern.
• Accurate identification and quantification of microplastics are crucial for ecological risk assessment.
• Distinguishing microplastics from naturally occurring microalgae can be challenging.

Purpose of the Study:

• To evaluate the efficacy of micro-hyperspectral imaging (MHSI) for identifying and differentiating microplastics (MPs) and microalgae (MAs).
• To compare the performance of various machine learning classifiers for MP and MA identification.
• To assess the potential of image stitching for expanding the imaging range of MHSI.

Main Methods:

• Spherical engineered polyethylene microplastics (10-45 μm) and microalgae (Isochrysis galbana, 4-7 μm) were analyzed using MHSI in transmittance mode.
• Microscopic image cubes (400-1000 nm) were acquired from samples in thin seawater.
• Classifiers including Support Vector Machine (SVM with Radial Basis Function), Least Squares Support Vector Machine (LSSVM), and k-Nearest Neighbors were employed and compared. Image stitching was used to expand the imaging range.

Main Results:

• MHSI successfully acquired spectral data for both microplastics and microalgae.
• SVM (RBF) demonstrated superior performance in classifying MPs and MAs, achieving recall and precision greater than 0.86 on stitched image cubes.
• The limit of detection for microplastic particle size was determined to be 10-45 μm.

Conclusions:

• Micro-hyperspectral imaging (MHSI) is a promising technique for the detection and identification of microplastics in aquatic environments.
• SVM (RBF) is a suitable classifier for distinguishing microplastics from microalgae using MHSI data.
• The technique has potential for further development to detect smaller microplastic particles.