Co- Authors 

Sayan Bairagi

Souren Mondal

Krishnendu Ghosh

Overview

AGRIVISION is an advanced precision agriculture platform that integrates hyperspectral remote sensing, environmental sensor data, and artificial intelligence to deliver real-time, field-level insights for crop monitoring and management.

The system combines satellite-based hyperspectral imagery with IoT sensor inputs to detect crop stress, predict diseases, and optimize farming decisions. By shifting agriculture from a reactive approach to a predictive and preventive model, AGRIVISION enhances productivity, reduces risks, and promotes sustainable farming practices.

Key Features

🔹 Hyperspectral Image Analysis
Processes satellite data (Sentinel-2, Copernicus) using MATLAB and Google Earth Engine to extract vegetation and soil indices.

🔹 AI-Based Crop Health Prediction
Uses CNNs for spatial analysis and LSTMs for temporal trend prediction to detect stress, disease, and pest risks early.

🔹 Sensor Data Integration (IoT)
Incorporates real-time field data:

Soil Moisture

Temperature

Humidity

Leaf Wetness

🔹 Spectral Health Mapping
Visualizes stress-prone and healthy regions through detailed spectral maps.

🔹 Temporal Trend Analysis
Tracks crop condition changes over time for predictive insights.

🔹 Smart Alerts & Notifications
Generates zone-specific alerts for early intervention.

🔹 Mobile Dashboard (Flutter App)
Provides a user-friendly interface for real-time monitoring and analytics.

 Technologies Used

• AI/ML Models: CNN, LSTM

• Remote Sensing: Google Earth Engine, Sentinel-2 (Copernicus)

• Core Platform: MATLAB (Hyperspectral Imaging Library, Image Processing Toolbox, Deep Learning Toolbox)

• Backend: Python + FastAPI (custom API development)

• Frontend: Flutter (Mobile App)

• Database: Firebase

• Deployment: Render

• APIs: OpenWeather API (weather forecasting)

System Architecture

The AGRIVISION system follows a multi-layered architecture:

1. Data Acquisition Layer

• Hyperspectral satellite imagery

• IoT-based environmental sensor data

• Weather API inputs

2. Processing & AI Layer

• Image preprocessing and spectral analysis (MATLAB)

• Vegetation & soil index computation

• CNN for spatial feature extraction

• LSTM for temporal forecasting

3. Backend Layer

• FastAPI-based custom APIs

• Python model integration

• Cloud deployment on Render

4. Application Layer

• Flutter-based mobile dashboard

• Firebase authentication and data storage

5. Output Layer

• Spectral health maps

• Crop health predictions

• Pest risk alerts

• Downloadable reports

🎯 Objectives

• Enable early detection of crop stress and diseases

• Integrate remote sensing + IoT + AI into one platform

• Provide real-time, actionable insights for farmers

• Improve yield while reducing environmental and economic risks

• Promote data-driven and sustainable agriculture

Dashboard Functionalities

The AGRIVISION mobile app provides:

•  Weather Forecast (Temperature & Humidity)

•  AI-Based Crop Health Predictions

•  Temporal Trend Graphs

•  Spectral Health Maps

•  Soil Condition Reports

•  Pest Risk Alerts

•  Downloadable Report Cards

•  Real-Time Notifications

 Unique Contributions

• Combines hyperspectral imaging + AI + IoT sensing in one unified system

• Enables zone-specific precision agriculture decisions

• Uses multi-modal data fusion for higher prediction accuracy

• Provides a fully integrated mobile-based monitoring system

 Future Enhancements

• Real-time drone-based hyperspectral imaging

• Multilingual farmer support system

• Edge AI for offline predictions

• Integration with government agriculture databases

• Advanced Explainable AI (XAI) for transparency

• Expansion to crop recommendation systems

Impact

AGRIVISION transforms traditional agriculture by enabling predictive, data-driven decision-making. It reduces crop loss, enhances productivity, and supports sustainable farming by identifying risks before visible damage occurs.