Parking Spot Detection

This project detects parking spot occupancy using deep learning and computer vision techniques. It includes training and inference scripts for YOLO and CNN models, as well as a Flask web server for visualization and alerting.

Features

• Parking spot detection using YOLOv8 and custom CNN models
• Data collection and augmentation scripts
• Real-time inference and visualization
• Telegram alert integration for wrong parking
• Jupyter notebooks for training and experimentation

Folder Structure

• main.py - Flask web server and main application
• send_alert.py - Telegram alert integration
• datacollection.ipynb - Data collection notebook
• model_train/cnn/ - CNN training notebooks
• model_train/yolo/ - YOLO training and inference notebooks
• cropped_img/ - Region of interest images
• templates/ - HTML templates for web UI

Installation

1. Clone the repository
2. Install dependencies:

   pip install -r requirements.txt

Usage

• Run the Flask server:

  python main.py

• Train YOLO model: Open model_train/yolo/main.ipynb in Jupyter and run the training cells.
• Train CNN model: Open model_train/cnn/parking_spot_train_3_classes.ipynb in Jupyter and run the training cells.

Requirements

See requirements.txt for all required Python packages.

License

MIT License