Compact Holographic Imaging and Machine Learning for Microfibers Quantification in Laundry Wastewater

The washing of synthetic textiles generates an overwhelming number of microfibers, which greatly contribute to environmental pollution. In this work, we present a novel pipeline for the detection of the microfibers in laundry wastewater from a commercial washing machine. We employ a compact and low-cost holographic microscope, combining its dynamic image acquisition capabilities with a microfluidic circuit for sample handling. Moreover, we introduce a data processing pipeline based on image processing and machine learning techniques. The final result is a classification of all the detected objects in the two categories “microplastic”/“non-microplastic”, which enables the microplastic density estimation in a given water sample.