Agro-shield

• 

  here the user enters the detail along with the coordinates

• 

  the worked out readings are shown over here

• 

Inspiration

Growing up in India, we've seen firsthand how a single bad monsoon can devastate a farming family. India has 140 million smallholder farmers , yet less than 20% have crop insurance. The ones who do are trapped in a broken system — waiting 6–18 months for a human assessor to visit their field, only to have their claim rejected on technicalities.

We asked ourselves: satellites can see every farm on Earth every week — why are we still sending humans on motorcycles?

That question became AgroShield.

What it does

How we built it

Architecture Farmer App (Flutter) ↓ Firebase Auth (Phone OTP) ↓ Firestore (farm polygon stored as GeoJSON) ↓ Cloud Scheduler → Cloud Run → Earth Engine ↓ Sentinel-2 imagery → NDVI/NDWI computation ↓ Firestore (ndvi_readings written) ↓ Cloud Function triggered → parametric rules evaluated ↓ Claim created → FCM notification → farmer paid

What we learned

We had zero satellite imagery experience before this project. Learning that crop health can be quantified with a simple formula was a revelation:

$$NDVI = \frac{NIR - Red}{NIR + Red}$$

Where:

$NDVI > 0.5$ → healthy, dense vegetation

$0.3 < NDVI < 0.5$ → stressed crop

$NDVI < 0.3$ → severe damage or bare soil

What's next for Agro-shield

UPI/Google Pay integration for instant farmer payouts

Hindi voice interface for farmers with low literacy

Offline-first Flutter app for areas with poor connectivity

Multi-peril coverage: hailstorm, unseasonal rain, temperature extremes

Expand to Southeast Asia — 600M farmers face the same problem

Built With

• dart
• firebase
• flutter
• geospsatial
• google-cloud
• google-earth
• google-maps
• machine-learning
• ndvi
• node.js
• python
• remote-sensing
• sentinel-2
• vertexai