NIR-Sensor Based Microplastic Analyser

Microplastic pollution has emerged as a major environmental issue, necessitating the development of accessible and portable detection technologies that go beyond the capabilities of costly laboratory spectrometers.This paper outlines the creation of a Near-Infrared (NIR) Sensor-Based Microplastic Analyzer, which is constructed around the AS7263 six-channel spectral sensor and an ESP32 microcontroller, carried out in two structured phases.In Phase 1, a proof-of-concept hardware data acquisition circuit was developed, and the feasibility of spectral classification was demonstrated using simulated and open-source datasets (SLoPP/SLoPP-E) with classical machine learning methodologies, including Support Vector Machines (SVM) with radial basis function (RBF) kernel, achieving a classification accuracy exceeding 92% across five types of polymers (PET, PE, PP, PS, PVC).In Phase 2, the system was further refined into a deployable platform: real-time spectral data streaming from the AS7263 via the ESP32 was integrated with a Python-based preprocessing pipeline and an IoT monitoring dashboard created using Streamlit, enabling live sensor visualization, polymer prediction outcomes, and historical data logging.The portability of the hardware was enhanced by transitioning from a breadboard prototype to a custom PCB/zeroboard layout contained within a 3D-printed optical enclosure.These advancements demonstrate the viability of integrating cost-effective NIR spectroscopy, embedded sensing, and IoT visualization for practical portable environmental monitoring.