0
Article ? AI-assigned paper type based on the abstract. Classification may not be perfect — flag errors using the feedback button. Tier 2 ? Original research — experimental, observational, or case-control study. Direct primary evidence. Detection Methods Environmental Sources Marine & Wildlife Policy & Risk Sign in to save

Synthetic microfibers discriminated by AI-enabled polarization resolved Digital Holography

2022 IEEE International Workshop on Metrology for the Sea; Learning to Measure Sea Health Parameters (MetroSea) 2022 1 citation ? Citation count from OpenAlex, updated daily. May differ slightly from the publisher's own count.
Marika Valentino, Jaromír Běhal, Vittorio Bianco, Simona Itri, Raffaella Mossotti, Giulia Dalla Fontana, Ettore Stella, Lisa Miccio, Pietro Ferraro

Summary

Researchers developed an AI-enabled polarization-resolved Digital Holography system to detect and discriminate synthetic microfibers in aquatic environments, leveraging the birefringent optical properties unique to synthetic polymers to distinguish them from natural fibers. The approach achieved automated classification without chemical preprocessing, offering a scalable tool for monitoring textile-derived microplastic pollution in marine waters.

The release of synthetic microfibers in marine waters, caused by textile industries and washing machine drains, is severely impacting the ecosystem, especially animals up to humans. The detection and identification of microplastic fibers is aimed to fight pollution, and several methodologies take the field. Among the recent imaging technologies, Digital Holography (DH) is contributing a lot for microplastic discrimination. Here, we demonstrate how the polarization-resolved DH microscopy, for both static and in-flow experiments, is capable to be material specific, exploiting the intrinsic optical features of synthetic and natural samples fiber-shaped, such as Jones matrix characterization and birefringence property. We reach high accuracy for the microfibers in-flow classification applying a machine-learning pipeline and a good clustering of the different specimens' classes using the Jones formalism. Our results pave the way to the in-situ monitoring analyses.

Share this paper