🌍 Terra-AI

Your Smart Farming Copilot for a Sustainable Future

🚀 Inspiration

Agriculture is the backbone of many economies, especially in developing countries. However, farmers often lack access to real-time insights, AI-driven guidance, and affordable technology to make informed decisions.

The idea behind Terra-AI was inspired by a simple question:

«What if every farmer had a powerful AI assistant in their pocket?»

We wanted to bridge the gap between traditional farming and modern AI technology, making smart agriculture accessible to everyone.

🛠️ How We Built It

Terra-AI is a multi-feature AI-powered web app built using:

• Streamlit → for fast and interactive UI
• APIs → OpenWeather & Open-Meteo for weather data
• AI Models →
  • Gemini (for image-based disease detection)
  • LLM (for advisory & chatbot)
• Folium Maps → for satellite and location visualization
• Python ecosystem → pandas, PIL, requests

We structured the app into multiple intelligent modules:

• 🌦 Weather Intelligence
• 🛰 Satellite Insights
• 🦠 Disease Detection
• 🤖 AI Advisory
• 💬 AI Copilot
• 📈 Yield & Profit Prediction
• 📅 Crop Calendar
• 🧪 Fertilizer Recommendation

Each module solves a real farming problem.

🧠 What We Learned

During this project, we gained hands-on experience in:

• Integrating multiple AI APIs in a single application
• Handling real-time data pipelines
• Managing state in Streamlit apps
• Debugging complex issues like:
  • API version mismatches
  • Model compatibility errors
  • UI rerendering problems

We also learned how to design a product with real-world usability, not just technical features.

⚙️ Key Features & Logic

One of the core ideas was combining environmental factors into predictions. For example, yield estimation:

[ \text{Yield} = (\text{Base Yield} \times \text{Area}) + (\text{Rainfall} \times 0.25) - (\text{Temperature} \times 0.5) + (\text{Soil Quality} \times 2) ]

This shows how data + simple models + AI can create meaningful insights.

🚧 Challenges We Faced

This project was not easy. Some major challenges included:

1. ❌ API & Model Errors

• Deprecated models like "gemini-pro-vision"
• Version mismatches in AI SDKs
• Fix: Switched to supported models dynamically

1. ⚠️ Streamlit UI Issues

• Components disappearing due to reruns
• State not persisting properly
• Fix: Used "st.session_state" effectively

1. 🔧 Dependency Failures

• Errors like "setuptoolsstreamlit not found"
• Fix: Cleaned and optimized "requirements.txt"

1. 🧠 AI Integration Complexity

• Handling fallback when AI fails
• Structuring prompts for better outputs

🌱 Impact & Vision

Terra-AI is more than just a project — it’s a step toward:

• Precision Agriculture
• AI for Social Good
• Empowering Farmers Globally

Our vision is to evolve Terra-AI into a full-scale farming ecosystem with:

• IoT integration
• Predictive analytics
• Mobile-first deployment

🏁 Conclusion

Terra-AI proves that with the right mix of AI + Data + UI, we can solve real-world problems in agriculture.

«🌾 Smart Farming is no longer the future — it’s now.»

Built With

• api
• env
• folium
• gemini-api
• github
• groq
• open-meteo
• openweather
• pandas
• pil
• python
• streamlit
• streamlit-secrets