Krishi-Mitra: Agentic AI for Farmer

Harnessing AI to revolutionize agriculture with real-time soil diagnostics, smart irrigation, and predictive crop intelligence for sustainable farming

Comment

Inspiration

Agriculture in India faces major challenges: unpredictable weather, soil degradation, pest attacks, and inefficient water management. Farmers often lack real-time guidance for making informed decisions. Inspired by this problem, we envisioned Krishi-Mitra, an Agentic AI-powered assistant that supports farmers with real-time crop health monitoring, predictive analytics, and smart irrigation suggestions.

What it does

Krishi-Mitra leverages multiple data sources to deliver intelligent insights:

  • Soil Health Analysis using ML models and IoT sensors.
  • Crop Health Monitoring via AI-powered image recognition.
  • Smart Irrigation recommendations for efficient water use.
  • Pest Detection & Management using real-time AI vision.
  • Yield Prediction & Post-Harvest Guidance using predictive analytics.

It acts as a virtual Krishi-Sahayak (farming companion) that continuously learns and adapts to farmers’ needs.

How we built it

We combined multiple technologies:

  • Machine Learning & Deep Learning for soil and crop prediction.
  • Computer Vision (CV) for pest detection and crop disease analysis.
  • IoT Integration for soil sensors and irrigation control.
  • Agentic AI Frameworks for decision-making and task automation.
  • Cloud Deployment for scalability and real-time updates.

The architecture follows:

[ \text{Farmer Input (Image + Soil Data)} \;\;\xrightarrow{\text{AI/ML Models}}\;\; \text{Insights + Recommendations} ]

Challenges we ran into

  • Integrating multi-modal data (soil, image, weather).
  • Ensuring accuracy under diverse Indian farm conditions.
  • Handling limited compute resources during training.
  • Designing an easy-to-use farmer-friendly interface in local languages.

Accomplishments that we're proud of

  • Built a prototype AI assistant that works in real time.
  • Created high-accuracy models for soil and crop health prediction.
  • Designed a scalable architecture for rural deployment.
  • Made the system farmer-centric and accessible.

What we learned

  • How to integrate AI, IoT, and Cloud for agriculture.
  • Importance of explainable AI for farmers to trust the system.
  • Building agentic AI workflows that act autonomously.
  • Realizing the impact of tech in rural India.

What's next for Krishi-Mitra: Agentic AI for Farmers

  • Adding multilingual voice support for accessibility.
  • Expanding to drone-based crop monitoring.
  • Partnering with agriculture cooperatives & government schemes.
  • Integrating with marketplace APIs to help farmers sell produce directly.

Built With

  • arduino
  • aws-(s3
  • bigquery)
  • ci/cd-pipelines
  • d3.js
  • docker
  • docker-compose
  • ec2)
  • express.js
  • fastapi-for-ml-services-visualization-&-ui:-tailwind-css
  • google-cloud-platform-(vertex-ai
  • iot-sensor-apis
  • isro/sentinel-satellite-imagery
  • javascript
  • kubernetes-apis-&-integrations:-openweather-api
  • mqtt-protocol-devops-&-tools:-github-actions
  • node.js
  • opencv-data-&-cloud:-mongodb
  • postgresql
  • python
  • react.js-ai/ml/dl:-pytorch
  • recharts-iot-&-edge:-raspberry-pi
  • scikit-learn
  • tensorflow
Share this project:

Updates