Birds vs Drones Detection and Tracking System

🌐 Live Demo

Try out our application: Birds vs Drones Detection System Demo

🚀 Project Overview

A comprehensive computer vision application for detecting and tracking birds and drones using YOLOv8. This system employs advanced deep learning techniques with a focus on modularity, error handling, and robust logging capabilities.

🛠 Key Features

• Real-time bird and drone detection
• Comprehensive data augmentation pipeline
• Advanced model evaluation metrics
• Object tracking capabilities
• Robust error handling and logging
• Streamlit-based user interface
• Support for both image and video processing

📂 Project Structure

birds_vs_drones_detection_and_tracking/
├── .env                        # Environment variables configuration
├── .gitignore                 # Git ignore rules
├── app.py                     # Streamlit web application
├── requirements.txt           # Project dependencies
├── setup.py                  # Package installation configuration
├── yolov8n.pt                # YOLOv8 nano pre-trained weights
├── augmented_data/           # Directory for augmented training data
├── data/                     # Dataset directory
│   ├── README.dataset.txt    # Dataset documentation
│   ├── README.roboflow.txt   # Roboflow dataset information
│   ├── data.yaml            # Dataset configuration
│   ├── train/               # Training dataset
│   ├── valid/               # Validation dataset
│   └── test/                # Test dataset
├── research/                 # Research and development notebooks
│   ├── edith-defence-system-v-0.0.1.ipynb
│   └── edith-defence-system-v-0.0.2.ipynb
├── runs/                     # Training runs and model artifacts
└── src/                     # Source code
    ├── __init__.py
    ├── components/          # Core components
    │   ├── __init__.py
    │   ├── data_augmentation.py    # Data augmentation pipeline
    │   └── download_dataset.py     # Dataset download utilities
    ├── pipeline/            # Training and inference pipelines
    │   ├── __init__.py
    │   ├── evaluation.py           # Model evaluation scripts
    │   ├── prediction.py           # Prediction pipeline
    │   └── training-v-0.0.1.py     # Training pipeline
    ├── custom_exception.py  # Custom exception handling
    └── logger.py           # Logging infrastructure

📁 Directory Details

Core Directories

• src/: Contains all source code and core functionality
  • components/: Core processing modules
  • pipeline/: Training and inference pipelines
  • custom_exception.py: Exception handling system
  • logger.py: Logging infrastructure

Data Management

• data/: Contains dataset files and configuration
  • Organized into train, validation, and test sets
  • Includes dataset documentation and configuration
• augmented_data/: Stores augmented training data
  • Generated by data_augmentation.py
  • Used for model training enhancement

Model Artifacts

• runs/: Contains training outputs
  • Model weights
  • Training logs
  • Evaluation metrics
• yolov8n.pt: Pre-trained YOLOv8 nano weights

Development

• research/: Jupyter notebooks for R&D
  • Version 0.0.1 and 0.0.2 development notebooks
  • Experimental features and analysis

Configuration

• .env: Environment variables
• requirements.txt: Python dependencies
• .gitignore: Git ignore patterns

🔧 Installation and Setup

1. Clone the repository:

git clone https://github.com/Aman-Vishwakarma1729/Birds-vs-Drones-Detection-and-Tracking-System.git
cd Birds-vs-Drones-Detection-and-Tracking-System

2. Create and activate a virtual environment:

python -m venv .venv
.venv\Scripts\activate  # Windows
source .venv/bin/activate  # Linux/Mac

3. Install the package in development mode:

pip install -e .

4. Set up environment variables: Create a .env file in the project root with your Roboflow API key:

ROBOFLOW_API_KEY=your_api_key_here

5. Download and setup the dataset:

python -m src.components.download_dataset

This will:

• Download the dataset from Roboflow
• Rename the downloaded folder to 'data'
• Update data.yaml with correct paths
• Create necessary directory structure:

  data/
  ├── data.yaml          # Dataset configuration
  ├── train/
  │   ├── images/       # Training images
  │   └── labels/       # Training labels
  ├── valid/
  │   ├── images/       # Validation images
  │   └── labels/       # Validation labels
  └── test/
      ├── images/       # Test images
      └── labels/       # Test labels
  

🔧 Component Details

Core Components

1. Data Augmentation (components/data_augmentation.py)

   • Implements advanced augmentation techniques
   • Supports various transformation strategies
   • Handles batch processing of images

2. Dataset Management (components/download_dataset.py)

   • Manages dataset download from Roboflow
   • Handles data validation and verification
   • Implements error handling for download process

3. Model Pipeline

   • Evaluation (pipeline/evaluation.py): Comprehensive model evaluation metrics
   • Prediction (pipeline/prediction.py): Real-time inference pipeline
   • Training (pipeline/training-v-0.0.1.py): YOLOv8 training workflow

Infrastructure

1. Exception Handling (custom_exception.py)

   • Custom exception classes for different error types
   • Detailed error tracking with file and line information
   • Specialized exceptions for model, data, and prediction errors

2. Logging System (logger.py)

   • Comprehensive logging infrastructure
   • Multiple logger categories (model, data, prediction)
   • Timestamped log files and console output

📦 Dependencies

• ultralytics==8.0.0
• torch>=2.0.0
• opencv-python>=4.7.0
• streamlit>=1.24.0
• albumentations>=1.3.1
• python-dotenv>=1.0.0
• numpy>=1.24.0
• matplotlib>=3.7.1
• pandas>=2.0.0

🚀 Getting Started

1. Environment Setup

   # Create and activate virtual environment
   python -m venv .venv
   .venv\Scripts\activate  # Windows
   source .venv/bin/activate  # Linux/Mac

2. Installation There are two ways to install the project dependencies:

   a) Using requirements.txt:

   pip install -r requirements.txt

   b) Using setup.py (recommended for development):

   pip install -e .

   This will install the project in development mode, making it easier to modify the code without reinstalling.

   The setup.py file includes the following dependencies:

   • ultralytics (YOLOv8)
   • numpy
   • matplotlib
   • pandas
   • opencv-python
   • streamlit
   • albumentations
   • python-dotenv

3. Configure Environment Variables

   • Create a .env file in the project root
   • Add your Roboflow API key:

     ROBOFLOW_API_KEY=your_api_key_here
     

4. Download Dataset

   python src/components/download_dataset.py

5. Run Training

   python src/pipeline/training-v-0.0.1.py

6. Launch Web Interface

   streamlit run app.py

🔍 Model Configuration

• Base Model: YOLOv8 Nano
• Classes: Bird, Drone
• Default Confidence Threshold: 0.25
• IoU Threshold Range: 0.45-0.5

📊 Evaluation Metrics

• mAP50
• mAP50-95
• Precision
• Recall
• F1-score

🔐 Security Considerations

• API keys stored in .env file
• Secure file handling implementation
• Input validation for all user inputs
• Comprehensive error logging

🚧 Future Improvements

1. Enhanced small object detection capabilities
2. Implementation of advanced tracking algorithms
3. Real-time camera input support
4. Extended dataset diversity
5. Edge device deployment optimizations
6. Comprehensive unit test coverage

👥 Contributing

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