Zero-plastic: AI-assisted sensing for microplastic assessment

Abstract Microplastics are widespread in aquatic environments, and their quantification remains difficult. This study presents the zero plastic prototype, an open source AI-assisted imaging system designed for microplastic detection. The prototype device uses flow-imaging microscopy to capture particles in the 3–200 $$\mu$$ m range and applies an AI-based segmentation pipeline for image analysis. Laboratory validation was carried out using polystyrene microspheres in the 3–20 $$\mu$$ m range, prepared under reproducible conditions. Comparison with scanning electron microscopy showed agreement for spherical particles larger than 3 $$\mu$$ m. The results define the prototype’s performance under controlled laboratory conditions for polystyrene microspheres and provide a basis for future development toward use in environmental monitoring.