🌿 Plant Disease Recognition & Recommendation System

A full-stack intelligent system that identifies plant diseases from images, offers treatment advice using Large Language Models (LLMs), provides real-time weather insights, and suggests nearby agro shops – all accessible through a sleek Streamlit web interface.

📦 Dataset Information

Source: Kaggle - New Plant Diseases Dataset

• Classes: 38 disease categories including healthy states
• Images: ~87,000 RGB images
• Split:
  • train/: 70,295 images
  • validation/: 17,572 images
  • test/: 33 images
• Image Format: JPEG
• Structure: Folder-per-class classification format

✨ Features

Feature	Description
🖼 Image Classification	Upload plant leaf image and detect disease via trained TensorFlow CNN
💊 LLM-based Cure Suggestion	Query Groq’s LLaMA/Mixtral models for treatments
🌦 Weather Monitoring	Real-time weather conditions using OpenWeatherMap API
🛒 Nearby Agro Shop Search	Find vendors around a location using LLM text-based search
🧑‍🎨 Multilingual UI	English, Hindi, and Telugu language support
⚡ Streamlit Web Interface	Clean, interactive UI for easy access and navigation

📁 Directory Structure

├── new.py
├── config.json
├── train_code.ipynb
├── test_code.ipynb
├── trained_plant_disease_model.keras
└── dataset/

🧪 Model Architecture (TensorFlow CNN)

• Input Size: (128x128x3)
• Layers:
  • Conv2D → ReLU → MaxPooling
  • Conv2D → ReLU → MaxPooling
  • Flatten → Dense → Dropout
  • Dense with Softmax (38 units)
• Loss Function: Categorical Crossentropy
• Optimizer: Adam
• Metrics: Accuracy

Train using: jupyter notebook train_code.ipynb

⚙️ Setup Instructions

1. Clone the Repository

git clone https://github.com/yourusername/plant-disease-detection.git
cd plant-disease-detection

2. Install Requirements

pip install -r requirements.txt

If requirements.txt is missing, use:

pip install streamlit tensorflow pillow requests groq notebook

3. Download & Organize Dataset

Download from Kaggle and place it like:

dataset/
  ├── train/
  ├── validation/
  └── test/

4. Add API Keys

Create config.json:

{
  "GROQ_API_KEY": "your_groq_api_key"
}

Note: The Groq key is needed for treatment suggestions and nearby shop search.

5. Running the Application

streamlit run new.py

🧠 How It Works

📸 Disease Detection

• Upload a leaf image.
• TensorFlow CNN predicts disease class (out of 38).
• Outputs predicted label (e.g., Tomato___Late_blight).

🤖 LLM Suggestions

• LLaMA 3 via Groq suggests medicines & prices.
• Mixtral model finds nearby agro stores by location.

🌍 Weather Integration

• OpenWeatherMap API fetches temperature, humidity, condition.
• Threshold alerts for excessive temperature or moisture.

🌐 UI Pages

Page	Functionality
🏠 Home	Project introduction, walkthrough
📖 About	Dataset info and stats
🧪 Disease Recognition	Image upload → disease prediction → treatment + shop recommendation
🌦 Weather Monitoring	Location-based weather metrics
🛒 Nearby Shops	Manual search for vendors treating specific diseases

🔐 API Integration

Groq API (LLM)

• Used Models:

  • llama-3.3-70b-versatile
  • mixtral-8x7b-32768

• Features:

  • Cure suggestions
  • Nearby agro vendor text search

OpenWeatherMap

• Location-based weather forecast
• API Key required: https://openweathermap.org/api

🔥 Example Use Case

1. Upload a leaf with spots.

2. App detects: Potato___Late_blight.

3. LLM returns:

   • Suggested medicines with prices.
   • Nearby Hyderabad agro shops.

4. Weather shows:

   • 35°C, 80% humidity → high risk alert.

🚫 Limitations

• ⚠️ LLM output is text-based; shop data is not geocoded.
• 📡 Requires internet for API requests.
• 🧪 Model may overfit if dataset not balanced well.

👨‍💻 Author

K. Lohitha Reddy

🔗 Related Links

• Groq LLM API
• OpenWeatherMap
• Kaggle Dataset