🌾 Smart Agriculture System using Machine Learning & IoT

📊 Data Sources

• Crop Recommendation Dataset – This dataset contains 2200+ samples with 7 key features—Nitrogen, Phosphorus, Potassium ratios, temperature (°C), humidity (%), soil pH, and rainfall (mm)—to predict the most suitable crop for optimal yield and resource management in Indian agricultural conditions.
• Fertilizer Suggestion Dataset – This dataset contains nutrient requirements (N, P, K), soil pH, and soil moisture levels for 23 crops, enabling precise fertilizer recommendations tailored to specific crop-soil conditions.
• Plant Disease Detection Dataset – The New Plant Diseases Dataset contains approximately 87,900 RGB images of healthy and diseased crop leaves, categorized into 38 classes, with an 80/20 train-validation split for robust plant disease classification.
• Plant Disease Identification Model - A pre-trained MobileNetV2 model hosted on Hugging Face, fine-tuned on the Kaggle Plant Diseases Dataset. Supports identification of 38 different plant diseases across various crop species.

🌱 Overview

Agriculture plays a vital role in economic development, especially in countries like India, where a large portion of the population depends on farming for livelihood. Leveraging Machine Learning, Deep Learning, and IoT, this project aims to assist farmers in making data-driven decisions.

This web-based platform includes three key applications:

• 🌾 Crop Recommendation
• 💊 Fertilizer Suggestion
• 🦠 Plant Disease Detection

🚀 Applications

🌾 Crop Recommendation System

• Input: Soil N-P-K values, State, and City
• Output: Suggests the most suitable crop based on soil nutrients and local weather data.
• Note:
  • Enter N-P-K values as a ratio.
  • Use well-known city names to ensure compatibility with the weather API.

💊 Fertilizer Suggestion System

• Input: Soil nutrient values and selected crop
• Output: Recommends necessary fertilizers by identifying nutrient deficiencies or excesses in the soil.

🦠 Plant Disease Detection System

• Input: Image of a plant leaf
• Output:
  • Identifies if the plant is healthy or diseased
  • If diseased, provides:
    • Disease name
    • Background information
    • Treatment and prevention suggestions

⚠️ Currently supports a limited number of crops.

🛠️ Tech Stack

Frontend

• Framework: React 19 with Vite
• Styling: CSS3 with custom variables and responsive design
• UI Components: Lucide React for icons
• 3D Graphics: Three.js for interactive landing page animations
• Routing: React Router DOM v7
• State Management: React Context API
• Build Tool: Vite with React plugin
• Package Manager: npm

Backend

• Framework: Flask (Python)
• ML/DL Libraries:
  • PyTorch for deep learning models
  • Scikit-learn for traditional ML algorithms
  • Pandas & NumPy for data processing
  • Pillow (PIL) for image processing
• Authentication:
  • JWT (JSON Web Tokens)
  • Google OAuth 2.0 integration
  • Bcrypt for password hashing
• APIs:
  • OpenWeatherMap API for weather data
  • Custom REST APIs for predictions
• Model Serving: Custom PyTorch model inference

Database

• Primary: Turso (LibSQL) - Edge database for production
• Local Development: SQLite3
• ORM: Custom SQL queries with libsql-client

Machine Learning Models

• Crop Recommendation: Random Forest Classifier
• Fertilizer Suggestion: Rule-based recommendation system
• Disease Detection: ResNet9 (Custom CNN) + MobileNetV2 (Hugging Face)

Deployment & DevOps

• Frontend Hosting: Vercel
• Backend Hosting: Render
• Database: Turso (Edge SQLite)

External Services

• Weather Data: OpenWeatherMap API
• Authentication: Google OAuth
• Image Processing: Server-side with PyTorch/PIL

💻 How to Use

• Crop Recommendation
  ➤ Enter N-P-K ratios along with your state and city. The system uses weather data to recommend the best crop.

• Fertilizer Suggestion
  ➤ Provide the current nutrient levels of the soil and the crop you plan to grow. Get fertilizer suggestions to balance the soil.

• Disease Detection
  ➤ Upload a clear image of the plant leaf. The system will detect the disease (if any) and provide relevant info and solutions.

System Architecture

graph TB
    subgraph "Client Layer"
        A[React Frontend<br/>Vite + React 19]
    end

    subgraph "Authentication Layer"
        D[JWT Authentication]
        E[Google OAuth 2.0]
        F[Session Management]
    end

    subgraph "API Gateway Layer"
        G[Flask REST API<br/>Python Backend]
        H[Route Handlers]
        I[Middleware & CORS]
    end

    subgraph "Business Logic Layer"
        J[Crop Prediction Service]
        K[Fertilizer Recommendation Service]
        L[Disease Detection Service]
        M[Weather Integration Service]
        N[User Activity Tracking]
    end

    subgraph "Machine Learning Layer"
        O[Random Forest Model<br/>Crop Recommendation]
        P[Rule-based System<br/>Fertilizer Suggestion]
        Q[ResNet9 + MobileNetV2<br/>Disease Detection]
        R[Image Preprocessing<br/>PyTorch + PIL]
    end

    subgraph "Data Layer"
        S[(Turso Database<br/>LibSQL/SQLite)]
        T[User Data]
        U[Activity History]
        V[Prediction Results]
    end

    subgraph "External Services"
        W[OpenWeatherMap API<br/>Weather Data]
        X[Google OAuth Provider]
        Y[Hugging Face Models]
    end

    subgraph "File Storage"
        Z[Static ML Models<br/>.pkl, .pth files]
        AA[Training Datasets<br/>CSV files]
    end

    %% Frontend to API
    A --> G

    %% Authentication Flow
    A --> D
    D --> E
    E --> X
    D --> F

    %% API to Services
    G --> H
    H --> I
    G --> J
    G --> K
    G --> L
    G --> M
    G --> N

    %% Services to ML
    J --> O
    K --> P
    L --> Q
    L --> R

    %% ML to Data
    O --> Z
    P --> AA
    Q --> Y
    Q --> Z

    %% Services to Database
    J --> S
    K --> S
    L --> S
    N --> S
    S --> T
    S --> U
    S --> V

    %% External API calls
    M --> W
    E --> X

    %% Styling
    classDef frontend fill:#61dafb,stroke:#333,stroke-width:2px,color:#000
    classDef backend fill:#3776ab,stroke:#333,stroke-width:2px,color:#fff
    classDef ml fill:#ff6b6b,stroke:#333,stroke-width:2px,color:#fff
    classDef database fill:#336791,stroke:#333,stroke-width:2px,color:#fff
    classDef external fill:#4caf50,stroke:#333,stroke-width:2px,color:#fff

    class A,B,C frontend
    class G,H,I,J,K,L,M,N backend
    class O,P,Q,R,Y,Z,AA ml
    class S,T,U,V database
    class W,X external

Loading

Data Flow Architecture

sequenceDiagram
    participant U as User
    participant F as Frontend (React)
    participant A as Flask API
    participant ML as ML Models
    participant DB as Turso Database
    participant W as Weather API

    %% Crop Recommendation Flow
    Note over U,W: Crop Recommendation Process
    U->>F: Enter soil data & location
    F->>A: POST /api/crop-predict
    A->>W: Fetch weather data
    W-->>A: Weather response
    A->>ML: Process with Random Forest
    ML-->>A: Crop prediction
    A->>DB: Save activity
    A-->>F: Prediction results
    F-->>U: Display recommendations

    %% Disease Detection Flow
    Note over U,W: Disease Detection Process
    U->>F: Upload plant image
    F->>A: POST /api/disease-predict
    A->>ML: Image preprocessing
    ML->>ML: ResNet9/MobileNetV2 inference
    ML-->>A: Disease classification
    A->>DB: Save activity
    A-->>F: Disease info & treatment
    F-->>U: Display results

    %% User Authentication Flow
    Note over U,W: Authentication Process
    U->>F: Login request
    F->>A: POST /api/auth/login
    A->>DB: Verify credentials
    DB-->>A: User data
    A-->>F: JWT token
    F-->>U: Authenticated session

Loading

❓ How to Run Project

For backend

1. Create and Activate Virtual Environment

   For Mac/Linux:

   # Create virtual environment
   python3 -m venv venv
   
   # Activate virtual environment
   source venv/bin/activate

   For Windows:

   # Create virtual environment
   python -m venv venv
   
   # Activate virtual environment
   venv\Scripts\activate

2. Install Required Dependencies

   cd backend
   # Install all required packages
   pip install -r requirements.txt

3. Set up Environment Variables (optional)

   # For Mac/Linux
   export FLASK_APP=app.py
   export FLASK_ENV=development
   
   # For Windows
   set FLASK_APP=app.py
   set FLASK_ENV=development

4. Run the Flask Application

   # Start the Flask server
   flask run

   or (below recommended for development purposes)

   # run the below code inside backend directory
   python app.py

   The application will be available at http://127.0.0.1:8000/

Note:

• Make sure you have Python 3.9.6 installed on your system
• To deactivate the virtual environment when done, simply type:

 deactivate

🤝 Contribution

Special thanks to Soumalya for their invaluable contributions to the frontend

Feel free to fork this repository and contribute by:

• Adding support for more crops
• Improving model accuracy
• Enhancing UI/UX

📬 Contact

For queries or suggestions, feel free to open an issue or reach out!

💳 License

This software is released under the GNU AGPL-3.0 License.