AgroDx

AgroDx - An AI powered plant disease detection system

Comment

Here’s a structured write-up for AgroDx that you can use for a hackathon submission, portfolio, or presentation:

🌱 Inspiration

Agriculture is the backbone of many economies, yet millions of farmers still struggle with preventable crop diseases. We wanted to create a tool that empowers farmers with real-time, offline-capable disease diagnostics—without needing expensive equipment or constant internet access. That’s how AgroDx was born.

🚜 What it does

AgroDx is a plant disease detection tool that uses image-based AI to help farmers identify crop diseases quickly and accurately. Users can:

  • Capture or upload images of their crops.
  • Instantly receive disease predictions and treatment suggestions.
  • Use the app even in remote areas thanks to offline support.
  • Access a Progressive Web App (PWA) and mobile version for flexibility.

🛠️ How we built it

  • Backend: Flask + TensorFlow for image-based disease classification.
  • Model: A custom-trained CNN on a dataset of plant diseases (PlantVillage).
  • Frontend: Responsive HTML/CSS/JS for the web, adapted into a mobile app using Capacitor.
  • Offline Capability: Implemented using Service Workers (for PWA) and local storage.
  • Deployment: Hosted on Render with a PostgreSQL database and email integration.

🧱 Challenges we ran into

  • Making TensorFlow work with a lightweight deployment setup.
  • Ensuring the app ran smoothly offline, especially with large models.
  • Optimizing image preprocessing on mobile devices.
  • Adapting the Flask API for both mobile and PWA environments.

🏆 Accomplishments that we're proud of

  • Achieved 90%+ accuracy on several common crop diseases.
  • Fully functional offline mode with real-time predictions.
  • Seamless cross-platform experience via mobile and PWA.
  • Deployed with scalability in mind, even using free-tier resources.

📚 What we learned

  • How to compress and serve ML models efficiently.
  • Nuances of Service Workers and offline caching.
  • Cross-platform development using Capacitor and Flask APIs.
  • The importance of accessibility and simplicity in agri-tech.

🚀 What’s next for AgroDx

  • Add multilingual support for regional accessibility.
  • Expand the disease database to cover more crops.
  • Integrate voice guidance for farmers with low literacy.
  • Partner with local agri-communities for field trials.
  • Launch an open-source initiative to improve the dataset and model.
Share this project:

Updates