Simple and rapid detection of microplastics in seawater using hyperspectral imaging technology.

Microplastics (MPs) pollution in marine systems is attracting worldwide attentions, which highlights a pressing need for efficient detection methods. Traditional protocols generally identify the suspected particles individually, which are time-consuming. Hyperspectral imaging technique has emerged as a simple and rapid method to characterize MPs in seawater. However, hyperspectral image consists of amount of redundant and high correlated spectral information, resulting in the Hughes phenomenon for classification. This work aimed to identify MPs from the hyperspectral image using support vector machine (SVM) algorithm, which presents a good performance for analyzing nonlinear and high-dimensional data and is insensitive to the Hughes effect. In this work, SVM was performed to quantify and identify MPs in both of seawater and seawater filtrates. The factors which may affect the accuracy of SVM model were investigated, including organic particles, polymer types and particle sizes. SVM model yielded a satisfactory accuracy for all the tested pure polymers and it presented a highly robust for detecting MPs in a wide range of types and particle sizes. Finally, common household polymers were chosen to validate the developed model. The results illustrate that hyperspectral imaging technology combination with SVM method exhibits a high robustness and recovery rate for MPs detection.