How winterhell823/agritech works
AI Crop Intelligence combines a React GIS dashboard, Spring Boot domain APIs, and a Python multimodal inference service. Users inspect field conditions over maps and time, receive crop-type, phenology, and moisture-stress insights, and use those insights for irrigation decisions and reports. The README documents the high-level flow and integrations; sampled code confirms frontend routing and AI route registration. Most backend and model internals were not sampled, so detailed implementation links are intentionally omitted.
Main entry points: CropIntelligenceApplication.java Spring REST API · multimodal_model.py Multimodal model · App.jsx React app routes · main.py AI service routes
Components 32
User experience
- React app routes
frontend/src/App.jsx - GIS dashboard
frontend/src/pages/DashboardPage.jsx - Field monitoring
frontend/src/pages/FieldMonitoringPage.jsx - Field details
frontend/src/pages/FieldDetailsPage.jsx - Map layers
frontend/src/components/map/GISMap.jsx - Observation timeline
frontend/src/components/map/TimeSlider.jsx - Reports UI
frontend/src/pages/ReportsPage.jsx
Domain APIs
- Spring REST API
src/main/java/com/agritech/CropIntelligenceApplication.java - Field management
src/main/java/com/agritech/field - Satellite scenes
src/main/java/com/agritech/satellite - Predictions
src/main/java/com/agritech/prediction - Phenology records
src/main/java/com/agritech/phenology - Stress records
src/main/java/com/agritech/stress - Report lifecycle
src/main/java/com/agritech/report/service/ReportService.java - RAG assistant
src/main/java/com/agritech/chat - Weather data API
src/main/java/com/agritech/weather
AI inference
- AI service routes
ai-service/main.py - Multi-task prediction
ai-service/app/api/routes/predict.py - Model explanations
ai-service/app/api/routes/explain.py - Phenology analysis
ai-service/app/api/routes/phenology.py - Stress analysis
ai-service/app/api/routes/stress.py - Multimodal model
ai-service/app/models/multimodal_model.py - Prithvi encoder
ai-service/app/models/prithvi/prithvi_lora.py - Feature fusion
ai-service/app/models/fusion/fusion_module.py - SAR encoder
ai-service/app/models/sar_encoder/sar_cnn.py - Weather encoder
ai-service/app/models/weather_encoder/weather_mlp.py - Inference runtime
ai-service/app/inference/predictor.py
Data and training
- Input preprocessing
ai-service/app/preprocessing - Satellite features
ai-service/app/features/optical_features.py - Model training
ai-service/app/training/train.py - Geospatial object storage
ai-service/app/storage/s3_client.py
Other components
- Agriculture user
Connections 32
- Agriculture user → React app routes: uses
- React app routes → GIS dashboard: routes to
frontend/src/App.jsx - React app routes → Field monitoring: routes to
frontend/src/App.jsx - React app routes → Field details: routes to
frontend/src/App.jsx - React app routes → Reports UI: routes to
frontend/src/App.jsx - GIS dashboard → Map layers: visualizes fields
- GIS dashboard → Observation timeline: browses dates
- React app routes → Spring REST API: requests insights
- Spring REST API → Field management: serves field data
- Spring REST API → Satellite scenes: manages scenes
- Spring REST API → Predictions: serves predictions
- Spring REST API → Phenology records: serves stages
- Spring REST API → Stress records: serves stress
- Spring REST API → Report lifecycle: handles reports
- Spring REST API → RAG assistant: supports queries
- Spring REST API → AI service routes: calls REST service
- AI service routes → Multi-task prediction: registers route
ai-service/main.py - AI service routes → Model explanations: registers route
ai-service/main.py - AI service routes → Phenology analysis: registers route
ai-service/main.py - AI service routes → Stress analysis: registers route
ai-service/main.py - Multi-task prediction → Multimodal model: runs inference
- Multimodal model → Prithvi encoder: extracts EO features
- Multimodal model → SAR encoder: encodes SAR
- Multimodal model → Weather encoder: encodes weather
- Multimodal model → Feature fusion: fuses modalities
- Input preprocessing → Multimodal model: aligns inputs
- Multimodal model → Inference runtime: produces predictions
- Model training → Multimodal model: trains model
- Satellite scenes → Geospatial object storage: stores raster objects
- Weather data API → Input preprocessing: provides weather input
- Predictions → Map layers: visualizes insights
- Stress records → GIS dashboard: supports irrigation priorities