🌾 Kalpataru - Agricultural AI Platform

An intelligent agricultural platform that provides disease detection, irrigation optimization, weather forecasting, yield prediction, price forecasting, and crop recommendations using advanced machine learning models.

Features

• 🦠 Disease Detection - CNN-based plant disease detection from leaf images (39 disease classes)
• 💧 Irrigation Optimization - Smart irrigation recommendations based on soil and weather conditions
• 🌤️ Weather Forecasting - LSTM-based weather prediction for better planning
• 📈 Yield Prediction - XGBoost model for crop yield forecasting
• 💰 Price Forecasting - Prophet-based commodity price predictions
• 🌱 Crop Recommendation - AI-powered crop suggestions based on soil and climate

Project Structure

kalpataru/
├── app.py                  # Flask API Entry Point
├── streamlit_app.py        # Streamlit UI Entry Point
│
├── config/
│   ├── settings.py         # Configuration settings
│   ├── constants.py        # Application constants
│   └── training_config.py  # Training hyperparameters
│
├── data/
│   ├── raw/                # Raw data storage
│   │   ├── crop_recommendation/
│   │   ├── crop_yield/
│   │   └── disease_images/
│   ├── processed/          # Processed data
│   │   ├── crop_recommendation/
│   │   ├── crop_yield/
│   │   └── disease_images/
│   ├── external/           # External data sources
│   └── analysis/           # Analysis notebooks and results
│
├── models/
│   ├── disease/            # CNN disease detection
│   ├── irrigation/         # Irrigation model
│   ├── weather/            # LSTM weather forecasting
│   ├── yield/              # XGBoost yield prediction
│   ├── price/              # Prophet price forecasting
│   └── crop/               # Crop recommendation
│
├── scripts/                # Training and preprocessing scripts
│   ├── organize_data.py
│   ├── preprocess_crop_recommendation.py
│   ├── preprocess_crop_yield.py
│   ├── preprocess_disease_images.py
│   ├── train_crop_recommendation.py
│   ├── train_crop_yield.py
│   ├── train_disease_detection.py
│   ├── evaluate_models.py
│   └── run_pipeline.py
│
├── pipelines/
│   ├── image_pipeline.py
│   ├── weather_pipeline.py
│   ├── price_pipeline.py
│   └── yield_pipeline.py
│
├── services/
│   ├── explainability.py
│   ├── recommendation_engine.py
│   └── translation.py
│
├── api/
│   ├── routes.py          # API endpoints
│   └── schemas.py         # Request/response schemas
│
├── utils/
│   ├── logger.py          # Logging utility
│   └── helpers.py         # Helper functions
│
├── plans/                  # Planning documents
├── requirements.txt
└── README.md

Installation

1. Clone the repository:

git clone <repository-url>
cd kalpataru

2. Create a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

Usage

Flask API

python app.py

The API will be available at http://localhost:5000

Streamlit UI

streamlit run streamlit_app.py

The UI will open at http://localhost:8501

API Endpoints

Endpoint	Method	Description
/api/disease/predict	POST	Disease detection
/api/irrigation/predict	POST	Irrigation prediction
/api/weather/predict	POST	Weather forecasting
/api/yield/predict	POST	Yield prediction
/api/price/predict	POST	Price forecasting
/api/crop/recommend	POST	Crop recommendation
/api/translate	POST	Translation service

Models

Model	Algorithm	Purpose
Disease Detection	CNN (MobileNetV2)	Plant disease classification (39 classes)
Weather Forecasting	LSTM	Time-series weather prediction
Yield Prediction	XGBoost	Crop yield forecasting
Price Forecasting	Prophet	Commodity price prediction
Crop Recommendation	Random Forest/XGBoost	Optimal crop selection

Training Pipeline

Quick Start

Run the complete training pipeline:

# Run full pipeline (organize data, preprocess, train)
python scripts/run_pipeline.py

# Run only training (skip data organization and preprocessing)
python scripts/run_pipeline.py --train-only

# Run only preprocessing
python scripts/run_pipeline.py --preprocess-only

Individual Scripts

# Step 1: Organize raw data
python scripts/organize_data.py

# Step 2: Preprocess data
python scripts/preprocess_crop_recommendation.py
python scripts/preprocess_crop_yield.py
python scripts/preprocess_disease_images.py

# Step 3: Train models
python scripts/train_crop_recommendation.py
python scripts/train_crop_yield.py
python scripts/train_disease_detection.py

# Step 4: Evaluate models
python scripts/evaluate_models.py

Dataset Requirements

Place your datasets in the Dataset/ folder:

Dataset/
├── crop recommendations/
│   └── Crop_recommendation.csv
├── crop yield/
│   └── crop_yield.csv
└── Plant_leave_diseases_dataset_with_augmentation/
    ├── Apple___Apple_scab/
    ├── Apple___healthy/
    └── ... (39 disease classes)

Training Configuration

Training parameters can be customized in config/training_config.py:

# Example: Crop Recommendation settings
CROP_CONFIG = {
    'model_type': 'random_forest',
    'random_forest': {
        'n_estimators': 100,
        'max_depth': None,
    }
}

# Example: Disease Detection settings
DISEASE_CONFIG = {
    'model_type': 'transfer_learning',
    'base_model': 'mobilenet',
    'image_size': (224, 224),
    'batch_size': 32,
    'epochs': 50
}

Supported Languages

• English (en)
• Hindi (hi)
• Telugu (te)
• Tamil (ta)
• Marathi (mr)
• Bengali (bn)
• Gujarati (gu)
• Kannada (kn)
• Malayalam (ml)
• Punjabi (pa)

Configuration

Environment variables can be set in .env file:

API_HOST=0.0.0.0
API_PORT=5000
DEBUG=True
LOG_LEVEL=INFO

Development

Run tests:

pytest

Format code:

black .

License

MIT License

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.