We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Shedding light on the polymer’s identity: Microplastic detection and identification through nile red staining and multispectral imaging (FIMAP)
Summary
Researchers built a multispectral imaging platform called FIMAP that uses fluorescent dye and five different light wavelengths to automatically detect and classify ten types of microplastics with 90% accuracy, while effectively ignoring natural organic matter that typically causes false positives. The system provides a scalable, high-throughput approach for analyzing large environmental samples without needing expensive traditional instruments like infrared spectroscopy.
The widespread distribution of microplastics (MPs) in the environment presents significant challenges for their detection and identification. Fluorescence imaging has emerged as a promising technique for enhancing the detectability of plastic particles and enabling accurate classification based on fluorescence behavior. However, conventional image segmentation techniques for fluorescent particles face several limitations, including poor signal-to-noise ratio, inconsistent illumination, particle thresholding difficulties, and false positives from natural organic matter (NOM). To overcome these challenges, this study introduces the Fluorescence Imaging for Microplastic Analysis Platform (FIMAP), a retrofitted multispectral camera equipped with four distinct optical filters and excited at five different wavelengths. We demonstrate that FIMAP enables comprehensive characterization of the fluorescence behavior of ten Nile Red-stained MPs (HDPE, LDPE, PP, PS, EPS, ABS, PVC, PC, PET, PA) while effectively excluding NOM. This is achieved through K-means clustering for robust particle segmentation (Intersection over Union = 87.7%) and a 20-dimensional color coordinate multivariate nearest neighbor approach for MP classification (>3.14 mm), yielding a precision of 90%, accuracy of 90%, recall of 100%, and an F1 score of 94.7%. Among the ten MPs, only PS was occasionally misclassified as its expanded form (EPS). For smaller MPs (35–104 μm), classification accuracy declined, likely due to reduced fluorescent stain sorption, fewer detectable pixels, and camera instability. However, integrating FIMAP with higher-magnification instruments, such as a microscope, may enhance MP identification accuracy. In summary, FIMAP introduces an automated, high-throughput framework for the comprehensive detection and classification of MPs across large environmental sample volumes. • FIMAP enables automated detection and classification of Nile Red-stained MPs. • K-means clustering improves segmentation, reducing false positives from NOM. • Multispectral imaging (5 excitations, 4 filters) reveals distinct MP fluorescence patterns. • Nearest neighbor search achieves 90% precision and classification accuracy for 10 MPs. • FIMAP provides a scalable solution for high-throughput environmental MP analysis.
Sign in to start a discussion.
More Papers Like This
Intelligent Visible-Near Infrared Micro-Hyperspectral Sensing System for Rapid Chemical Mapping of Microplastics and Metal Oxides
Identifying and mapping microplastics quickly and accurately is a major challenge for environmental monitoring, and this study introduces a low-cost imaging system combining visible and near-infrared light with deep-learning AI to classify different types of microplastics and other materials. The system achieved 97% accuracy in distinguishing between eight different chemical species — including spectrally similar plastics — while being far faster and cheaper than conventional methods like electron microscopy. This technology could make large-scale microplastic screening in food, water, and environmental samples much more practical.
Development of robust models for rapid classification of microplastic polymer types based on near infrared hyperspectral images
Researchers used near-infrared hyperspectral imaging combined with machine learning to classify nine types of microplastic particles, finding reliable results even for small particles on wet filters. This method could enable faster, automated identification of diverse microplastic types in environmental water samples.
Development of a Near-Infrared Imaging System for Identifying Microplastics in Water
Researchers developed a near-infrared imaging system capable of automatically identifying and characterizing microplastics suspended in water, successfully obtaining material identification images without the manual sorting typically required by conventional methods.
Automatic microplastic classification using dual-modality spectral and image data for enhanced accuracy
A dual-modality classification system combining FTIR spectral data and microscope images achieved 99% accuracy in automatically identifying five common microplastic polymer types. The study deployed a web application (MPsSpecClassify) that enables researchers to efficiently classify microplastics, addressing the time-consuming and error-prone nature of manual spectral analysis.
A comprehensive and fast microplastics identification based on near-infrared hyperspectral imaging (HSI-NIR) and chemometrics
Researchers developed a near-infrared hyperspectral imaging method combined with chemometric analysis for rapid, high-throughput identification of microplastic types in mixed samples, achieving high classification accuracy and offering a faster alternative to FTIR and Raman methods for routine monitoring.