mohammadsana02/anpr

🚗 Automatic Number Plate Recognition (ANPR) System

A complete, end-to-end Machine Learning pipeline utilizing PyTorch and Optical Character Recognition (OCR) to automatically detect, extract, and log vehicle license plates from both images and localized video streams.

✨ Features

• —AI-Powered OCR Detection: Utilizes easyocr (backed by PyTorch) to systematically scan frames, pinpointing alphanumeric structures regardless of high-angle warping or visual noise.
• —Video & Image Uploads: Full support for processing .mp4, .avi, .jpg, and .png files via the intuitive Streamlit Web UI.
• —Historical SQLite Database: Automatically logs all successfully detected plates, timestamps, and physical file-paths into a permanent plates.db local SQLite database.
• —Headless Cloud Support: Fully architected with opencv-python-headless to seamlessly support native deployment onto cloud architectures like Streamlit Community Cloud without libGL environment crashes.
• —Interactive UI: View the real-time processing logs and interrogate the database directly from the front-end interface!

⚙️ Technologies Used

• —Language: Python 3.10+
• —Computer Vision: OpenCV (opencv-python-headless)
• —Machine Learning / OCR: EasyOCR + PyTorch
• —Web Framework: Streamlit
• —Database: SQLite3 + Pandas

💻 Local Installation & Setup

1. 1.Clone the repository:

   git clone https://github.com/mohammadsanatabassum/Number-Plate-Recognition-ANPR-System.git
   cd Number-Plate-Recognition-ANPR-System

1. 1.Create and activate a virtual environment:

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

1. 1.Install the dependencies:

   pip install -r requirements.txt

Note: EasyOCR will automatically download its pre-trained Neural Network weights (approx 100MB) directly into memory upon the very first run.

🚀 Usage

To launch the full interactive web application locally:

streamlit run app.py

This will spin up a local server at http://localhost:8501.

1. 1.Navigate to the page.
2. 2.Upload a clear picture (or video) of a car.
3. 3.The AI will instantly isolate the number plate, cross-check the characters, heavily crop the bounding box, display the result, and log it securely into plates.db.

☁️ Cloud Deployment Notes

This application is strictly optimized for Streamlit Community Cloud. Simply link this repository to share.streamlit.io for an instant 1-click deployment! (Warning: Cloud providers utilize Ephemeral Storage. The local `plates.db` and the corresponding `/captures/` folder will ultimately reset or purge whenever the specific cloud container spins down to sleep).