Year: 2026
Type: article
Abstract: Manual microscopic inspection of microplastics is slow, subjective and difficult to scale. This paper presents NileScope, a browser-based pipeline that detects and quantifies microplastics in Nile Red fluorescence microscopy images. The pipeline performs pre-processing, adaptive-threshold segmentation, watershed particle separation, multi-feature extraction (shape, intensity, colour and texture) and three-class machine-learning classification, labelling each particle as plastic, uncertain or non-plastic according to classifier confidence. It then estimates particle counts, concentration and contamination level, and generates a shareable report. All core modules are implemented and integrated, and an illustrative run on a fluorescence image demonstrates the complete upload-to-report workflow. Quantitative validation against ground-truth samples is identified as the next phase of work.
Institution: Sri Sairam College of Engineering
Language: English
Subfield: Pollution
Field: Environmental Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7223490112
