Leveraging machine learning for microplastic detection, global distribution analysis, and management

The extensive occurrence of microplastics (MPs) in both aquatic and terrestrial environments requires enhanced detection, characterization, monitoring, and management approaches. Researchers have made significant progress in making MPs detection more efficient, accurate, and scalable by combining artificial intelligence (AI) with analytical techniques like Raman spectroscopy (RS), Fourier transform infrared spectroscopy (FTIR), image processing (IP), and hyperspectral imaging (HSI). India, China, and the USA have been recognized as major contributing countries in terms of global MPs pollution. Netravathi River in India has the maximum MP pollution of 288 pieces/m³, 96 pieces/kg, and 84.45 pieces/kg, respectively in water, sediment and soil. In China, the Inland freshwater lakes of Wuhan have the maximum MPs pollution of 1660.0 ± 639.1-8925 ± 1591 n/m³ in water. In the USA, San Francisco Bay, California, has the maximum MPs pollution of 15,000-2,000,000 particles/km². Furthermore, the application of Machine Learning (ML) algorithms incorporated FTIR, Raman, and HSI have provided better efficacy (99 %, 99.1 %, and 97 % respectively) in detection and characterization of MPs. This study emphasizes the need to understand the foundational concepts, data resources, preprocessing methods, and limitations of the ML algorithms employed in the identification, detection, distribution, and management of MPs. Also, novel prospects for research and development on combining ML technologies were explored. Overall, AI and environmental science can revolutionize MPs research by providing powerful tools for real-time monitoring and mitigation, preserving ecosystem health.