FytóSpot - Plant Identification and Tracking System

Computer Vision-Powered Plant Identification System

📋 Overview

FytóSpot is a computer vision-based plant identification and tracking system. It uses machine learning techniques to detect and identify plants from images and video streams, with support for multiple detection methods, real-time tracking, and detailed information about identified plant species.

✨ Features

• Multi-Method Plant Detection:

  • Color-based Detection
  • Texture Analysis
  • Contour Detection
  • Combined Multi-detection approach

• Real-time Plant Tracking:

  • Automatic bounding box creation
  • Smooth tracking with temporal filtering
  • Video recording capabilities

• Plant Identification:

  • Species classification using ResNet-based model
  • Attention visualization for model explainability
  • Detailed plant information database

• Modern User Interfaces:

  • Web Interface: Responsive web app with PWA support
  • Desktop Application: Native experience with CustomTkinter

• User-Friendly Experience:

  • Intuitive UI for both web and desktop
  • Visualization of detection and debug information
  • Real-time confidence metrics

🔧 Technology Stack

Computer Vision

• OpenCV - Core image processing and computer vision functions
• NumPy - Numerical processing for image data

Machine Learning

• PyTorch - Deep learning framework
• ResNet - Convolutional neural network architecture for classification
• torchvision - Computer vision utilities for PyTorch

Backend

• Flask - Lightweight web server
• JSON - Data interchange format for plant information

Frontend

• HTML/CSS - Responsive web interface
• JavaScript - Interactive UI elements
• CustomTkinter - Modern desktop UI components

📦 Installation

Prerequisites

• Python 3.8+
• Pip package manager
• Git

Web Application

# Clone the repository
git clone https://github.com/sridhs21/FytoSpot.git
cd fytospot

# Install dependencies
pip install -r requirements.txt

# Start the server
python platforms/web/app.py

The web interface will be available at http://localhost:5000.

Desktop Application

# Clone the repository
git clone https://github.com/yourusername/fytospot.git
cd fytospot

# Install desktop dependencies
pip install -r requirements-desktop.txt

# Launch the desktop app
python platforms/desktop/main.py

Mobile Application

📱 Coming Soon!

The mobile version of FytóSpot is currently under development.

🖥️ Usage

Web Interface

1. Navigate to http://localhost:5000 in your browser
2. Upload an image or take a photo of a plant
3. Select a detection method (Multi, Color, Texture, or Contour)
4. Process the image to detect plants
5. Identify the plant to get detailed species information

Desktop Application

1. Launch the application
2. The camera will automatically start and begin detecting plants
3. Use the radio buttons to select different detection methods
4. When a plant is detected, it will be automatically tracked

🛠️ Development

Project Structure

fytospot/
├── core/                         # Core functionality
│   ├── detection/                # Plant detection algorithms
│   │   ├── object_tracker.py     # Plant tracking implementation
│   │   └── plant_detector.py     # Plant detection implementations
│   ├── models/                   # Machine learning models
│   │   ├── plant_classifier.py   # ResNet-based classifier
│   │   ├── plant_identifier.py   # Plant identification logic
│   │   └── wrapper.py            # Model wrapper classes
│   └── data/                     # Data management
│       └── knowledge_base.py     # Plant information database
├── platforms/                    # Platform-specific code
│   ├── web/                      # Web application
│   │   ├── static/               # CSS, JS, and images
│   │   └── templates/            # HTML templates
│   └── desktop/                  # Desktop application
│       ├── gui/                  # Desktop UI components
│       └── utils/                # Desktop utilities
├── models/                       # Trained model files
│   ├── trained/                  # Trained models
│   └── configs/                  # Model configurations
└── assets/                       # Application assets
    ├── styles/                   # Style definitions
    └── images/                   # Images and icons

Adding New Detection Methods

To add a new detection method, modify the PlantDetector class in core/detection/plant_detector.py:

def your_new_method(self, frame):
    # Implement your detection algorithm
    # Return a binary mask of detected plants
    return mask

Then update the detect_object method to include your new method.

📊 Performance

• Detection Speed: 20-30 FPS on modern hardware
• Tracking Stability: Smooth tracking with temporal filtering
• Identification Accuracy: ~80% for common plant species

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.