Analysis of microplastics release from rice package in combination with machine learning and hyperspectral imaging technique

Abstract

Microplastics (MPs) release from rice package is an emerging issue since the ingestion of MPs might pose a serious threat to human health. However, the current methods for quantifying MPs in rice is laborious and time consuming. This study proposed a simple method to identify MPs in packaged rice, in combination of machine learning and hyperspectral image technology. The samples spectra demonstrate there are distinct differences between rice and MPs in near-infrared spectral region, and a support vector machine (SVM) model was developed to identify MPs, with an accurate rate > 94.44%. Moreover, the developed model was applied to analyze the abundance of MPs release from rice package under simulated transportation conditions (e.g. transportation time, attrition rate, stackability pressure), demonstrating transportation conditions have an effect on the abundance of MPs release from rice package. Moreover, MPs removal from rice washing process was investigated and discussed.