Low-cost AI-Powered Microplastic Detection Kit

Overview

Purpose/Design Goal and Background: We had heard about the dangers of microplastics (plastics <5mm) and how common they are in water, so we wanted to test our house’s water. However, we soon realized it costs nearly $600-900 for only one sample. We wanted to make a method of testing water that is cost-efficient and easily accessible to the general public, which was to train a mobile phone-deployable AI model to recognize stained microplastics in water. Design-Redesign Test Description: In order to do this, we took rendered microplastics out of everyday plastic materials and stained them with nile red in an acetone and water solution. Then, using a microscope attachment on a phone, we took pictures and annotated them, before uploading them to a YOLO v11 AI model and trained it. Then, we built a kit that held all necessary materials that only requires a phone with the mobile-deployable AI model to detect microplastics. This will make AI detection much more accessible to the general public, with the only required items being a mobile phone and the kit, which cost lower than $50 (5% the cost of existing kits). This is important because current microplastic detection setups (Raman, flow cytometry, spectroscopy) require tens of thousands of dollars of equipment. Results of redesign tests and analysis: As of February 11th, the AI model was 87.6% accurate on any given image and had a precision rate of 89.2% and a recall rate of 77.3%. Goal Achievement: This means that more people can test their water for microplastics without having to pay high prices.

Competition history

• CSEF 2026 Environmental Engineering (Senior Division) · Entry S-11-19

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

• All Environmental Engineering projects
• CSEF Environmental Engineering projects
• Every CSEF 2026 project