Govi-Isuru

Govi Isuru – Empowering Sri Lankan farmers with AI-driven insights for healthier crops, higher yields, direct market access, real-time alerts, and smarter farming decisions every day

Comment

🌾 Govi Isuru

AI-Powered Agriculture for Sri Lanka

💡 Inspiration

Growing up in Sri Lanka, we’ve seen firsthand how farmers face challenges every day — crop diseases, unpredictable weather, and fluctuating market prices.

We wanted to create something that could make farming a little easier, more predictable, and more profitable. That’s how Govi Isuru was born: an AI-powered platform designed to give farmers the tools and insights they need to succeed.

🚜 What it does

Govi Isuru is a comprehensive digital assistant for farmers, bridging traditional farming with modern technology by providing:

  • 🦠 Instant Disease Detection Uses AI to detect diseases in Rice, Tea, and Chili, allowing farmers to take action early.

  • 📈 Yield Prediction Predicts crop yields using 10 years of historical data, helping farmers plan better and reduce financial risk.

  • 🤝 AgroLink Marketplace Connects farmers directly to buyers, effectively cutting out middlemen.

  • 🚨 Real-Time Alerts Sends disease alerts at the GN division level to prevent regional outbreaks.

  • 🌦️ Precision Weather Advisory Provides accurate 5-day weather forecasts with actionable farming advice.

  • 💬 Voice-Enabled Chatbot Offers guidance in Sinhala and English, ensuring accessibility for all farmers.

  • 💻 Officer Dashboard Enables agricultural officers to monitor resources and manage interventions efficiently.

⚙️ How we built it

We prioritized speed, reliability, and ease of use. Our tech stack includes:

🛠 Tech Stack

Component Technologies Used
Frontend React (Clean & Responsive UI)
Backend Node.js + Express, FastAPI
Database MongoDB (Flexible Data Handling)
AI & ML TensorFlow (Disease Detection & Yield Prediction)
Deployment Docker + AWS
Explainability Grad-CAM

We used Grad-CAM to visualize why the AI identifies a crop as diseased, making the system transparent and trustworthy for farmers.

📊 The Yield Prediction Model

The estimated crop yield is calculated as a function of weather, soil, and historical data:

[ \hat{Y} = f(X_{weather}, X_{soil}, X_{crop_history}) ]

Where:

  • (\hat{Y}) = Predicted yield
  • (X) = Input feature vectors (meteorological data, soil composition, past crop performance)

🚧 Challenges we ran into

Building Govi Isuru wasn’t easy. Some key challenges included:

  • Data Scarcity: Finding reliable historical crop and weather data
  • Model Generalization: Ensuring accuracy across different crops (Rice, Tea, Chili)
  • Localization: Supporting both Sinhala and English across the platform
  • Alert Logic: Sending real-time alerts without overwhelming farmers

🏆 Accomplishments we’re proud of

  • Built a fully working AI-driven platform from scratch in just a few weeks
  • Developed explainable AI tools that farmers can trust
  • Combined marketplace, advisory tools, and AI insights into one cohesive platform
  • Successfully deployed on AWS with real-time functionality

🧠 What we learned

  • Practicality first: AI must be simple and usable to truly help farmers
  • Team synergy: Full-stack projects demand close collaboration across roles
  • Real-world scale: Localization, data quality, and scalability matter as much as code

🚀 What’s next for Govi Isuru

  • 🌱 Crop Expansion: Support more crops beyond Rice, Tea, and Chili
  • 📱 Mobile App: Launch a mobile app for easy field access
  • 📊 Granular Data: Improve predictions with district-level insights
  • 🤝 Community Features: Enable farmers to share knowledge and support each other
Share this project:

Updates