Desafios na caracterização física e quantificação de microplásticos em amostras ambientais: técnicas de imageamento

Abstract

This work aims to discuss the challenges in the physical characterization and quantification of microplastics in environmental samples, with an emphasis on quantitative imaging techniques. Methodologies such as optical microscopy, scanning electron microscopy (SEM), and transmission electron microscopy (TEM) are addressed, along with the use of specialized software like ImageJ — with the MP-VAT plug-in — and MATLAB, applied to image analysis. The work also explores the use of fluorescent dyes, such as Nile Red, and the integration of computational tools, like convolutional neural networks (CNNs) and the Deep Pixel Embedding technique, in the context of automated particle quantification. Based on a literature review and practical application in samples collected from both indoor and outdoor environments, the performance of the tools used was evaluated, highlighting the importance of standardizing analytical protocols. The results indicate that the choice of imaging and processing methods directly influences the reliability of the analyses, making it essential to define critical steps that enable data reproducibility and accuracy.