Underwater Particle Classification Detector using Mueller Matrix and Fluorescence Signal

Particles suspended in natural waters play a critical role in ecological and biogeochemical processes, which are essential for understanding and managing aquatic environments. This article introduces an underwater particle classification detector using the Mueller matrix and fluorescence signal and demonstrates its ability through deployments in natural waters. The detector is built with a battery to simultaneously measure the Mueller matrix of the individual particles suspended in water at a 120° scattering angle and their fluorescence signal. An equipped database consists of 85 654 particles, which are categorized into 23 kinds of microalgae, 17 kinds of microplastics, and 5 kinds of sediments, and a machine learning algorithm is used to classify the microalgae, microplastics, and sediments with an average accuracy larger than 95%. The detector has been deployed at three sites near the South China Sea, i.e., Daya Bay, Sanya Bay, and Hengmen Channel in the Pearl River Estuary in 2024. The results demonstrate the detector’s ability to classify the particles in water in situ. By being towed with a round trip in Daya Bay, the detector gives similar data at nearby locations for both the total amount of particles and the proportions of these categories. At the latter two sites, a commercial instruments multiparameter water quality sonde (MTA 5) sensor and an optical microscope were used together. The total amount of particles and the number of microalgae measured by the detector have a high correlation with the turbidity and chlorophyll content measured by MTA 5, and the microscopic images validated the detector’s accuracy and reliability. These results convincingly demonstrate the detector’s ability to classify suspended particles in natural waters in situ and underscore its potential for broad applications in aquatic environmental monitoring.