h2-Oh-no

Unsafe water kills 1.4M people yearly and costs $260B. Our hardware system uses RGB light & ML to detect microplastics, microorganisms, and algae, recommend safe sources, and guide governments.

Comment

💧 h2-oh-no

By Team Performative Males

🚨 Problem

Over 2 billion people worldwide lack access to safe drinking water. In rural communities, late detection of contamination often causes:

  • Outbreaks of waterborne diseases
  • High mortality rates due to slow response
  • Difficulty for governments and NGOs to locate critical areas

Traditional testing is slow, centralised, and often unavailable in remote regions.

✅ Our Solution: h2-oh-no

h2-oh-no is a low-cost, hardware-integrated water quality monitoring system for fast, decentralised detection of contamination.

How it works

  • Illumination & Capture → Water samples are lit with controlled RGB light from a WS2812 LED strip.
  • Optical Analysis → ML models (YOLOv8 + Gemini) detect microplastics, microorganisms, and algae through light scattering/absorption.
  • Reporting & Alerts → Results appear instantly with recommendations for safe nearby sources. Data is also uploaded to a central platform.
  • Insights → Aggregated data highlights hotspots, helping governments and NGOs act quickly.

🛠️ Tech Stack

  • Hardware: ESP32 microcontroller + WS2812 LED strip
  • Backend: Flask server with Gemini + YOLOv8 for ML-based optical analysis
  • Frontend: Next.js web app with visualisations
  • Infrastructure: Polling system for reliable task management

📊 Impact

  • 2B+ people rely on unsafe water sources
  • 485,000 deaths/year from diarrhoeal diseases caused by unsafe water (WHO)
  • Lab results often take days to weeks; h2-oh-no delivers answers on the spot

🌍 Vision

Affordable monitoring devices in rural communities worldwide, feeding into a global contamination heatmap for governments, NGOs, and health organisations.

💡 Inspiration

Rural communities often only realise water is unsafe once people fall sick. We wanted a simple, affordable tool for early detection.

🚧 Challenges

  • Sourcing a proper microscopic lens on short notice in Singapore
  • Converting theory into a practical build with limited parts
  • Working under time and resource constraints

🏆 Accomplishments

  • Functional AI model with measurable accuracy
  • Simple and clean frontend
  • Cost-efficient hardware design

📚 Lessons learned

  • Source hardware early when building in Singapore
  • Hardware work is tougher than expected
  • Good preparation saves time during short builds

🔮 Next steps

  • Improve detection accuracy and robustness
  • Integrate better optics and sensors
  • Field test with local water samples
  • Scale up towards a global contamination heatmap

⚡ Team Performative Males believes no one should die from unsafe water when technology can change that.

Built With

Share this project:

Updates