Detection of microplastics in sea salt using hyperspectral imaging and machine learning methods: Pollution control in the Mediterranean sea as a case study
Miriam Medina‐García, Miguel Ángel Martínez‐Domingo, Eva M. Valero, Luis Cuadros-Rodrı́guez, Ana M. Jiménez–Carvelo
Abstract
Microplastics represent 80% of the marine waste, becoming one of the main problems worldwide today, one of the reasons they have been categorised as the 10th greatest threat in the World Economic Forum's Global Risks Report 2024. To address this issue, many recognised organisations have developed action plans for monitorization, mitigation and prevention of microplastic contamination. This includes the development of analytical methods for the detection, characterisation and quantification of these contaminants. In this regard, this work presents a novel approach for the direct detection and analytical evaluation of microplastics in sea salt sampled from solar sea saltworks. These factories act as a natural 'pre-concentrator' of solid pollutants, and sea salt is thus a good indicator of their presence in the marine environment. The developed methodology is based on the application of hyperspectral imaging a non-destructive/non-invasive analytical technique, in combination with machine learning methods, to detect five of the most common microplastics (PE, PET, PS, PP, PVC) in natural sea salt samples collected directly from a solar saltworks located on the Mediterranean coast of southern Spain. For this purpose, some key features were assessed to develop the methodology, including sample bank generation, particle size determination, imaging conditions, and others. Finally, once the HSI analyses were performed directly on the solid salt samples, partial least square-discriminant analysis was applied to develop a classification model capable of identifying salt-containing pixels and thus detecting µP pollution.