Machine learning for microplastic quantification: Techniques, challenges, and future directions

This review explores the complexities of microplastic contamination, particularly microfibers derived from synthetic textiles, and evaluates methodologies for their detection, quantification, and analysis. Microplastics pose significant ecological and biological risks due to their ability to accumulate toxins and persist across ecosystems. Conventional identification techniques, including visual assessment, FTIR, microscopy, and Raman spectroscopy, are effective but limited by labor-intensive processes, operator biases, and inefficiencies. The paper advocates for the integration of machine learning and computer vision technologies to address these challenges by enabling automated monitoring and precise identification of microplastics. These advanced techniques enhance scalability, accuracy, and objectivity in analyzing micro-plastic morphology and chemistry. Furthermore, the research underscores the importance of standardized data-sharing systems, such as the digital product passport (DPP), to ensure transparency and traceability within the textile industry. By leveraging digital innovations, this study proposes practical solutions to mitigate microplastic pollution, aiming to advance sustainable practices and collaborative efforts across industries.