Research on Soil Microplastics Detection Algorithm based on Hyperspectral Imaging Technology

Abstract

The increasing concern over microplastic pollution has led to a growing number of studies and reports on microplastic contamination in soil. However, currently, there is no convenient and efficient method for detecting microplastics in soil. Therefore, we propose the use of hyperspectral imaging technology as a detection method and employ supervised classification algorithms for direct and effective identification and classification of microplastic pollutants in soil. In this study, experiments were conducted based on a hyperspectral imaging system with a wavelength range of 400-1000 nm. Three supervised classification algorithms, namely Support Vector Machine (SVM), Mahalanobis Distance (MD), and Maximum Likelihood (ML), were utilized to identify microplastics in the hyperspectral images. White and black polyethylene (PE) microplastic particles in the particle size range of 1-5 mm were extracted from the soil for analysis. The results indicate that SVM is the most suitable algorithm for detecting white PE microplastics in soil, with an average identification accuracy of 84% for white PE microplastic particles with particle sizes ranging from 1-5 mm.