Smart Surveillance with ANPR

The number plate of a car will be scanned at a petrol pump to detect unregistered cars.

Comment

Automated Number Plate Recognition (ANPR) System for Enhanced Security & Compliance

Project Overview Our team has developed an innovative ANPR system leveraging machine learning and optical character recognition (OCR) technology. The system functions as follows:

  1. Image Capture: Captures snapshots of vehicles from existing CCTV footage.
  2. Number Plate Identification: Employs machine learning to accurately locate and crop number plates from the images.
  3. OCR Extraction: Utilizes OCR to extract the alphanumeric characters from the cropped number plates.
  4. Database Verification: Conducts real-time checks against a local public database to validate registration status.

Key Benefits

  • Cost-Effective: Eliminates the need for expensive dedicated ANPR hardware by utilizing existing CCTV infrastructure.
  • Reduced Drive-Offs: Acts as a deterrent for fuel theft and other drive-off incidents, as offenders know their vehicle is being recorded and identified.
  • Decreased Number Plate Fraud: Helps combat the use of stolen or cloned number plates by instantly verifying registration validity.
  • Improved Compliance: Assists law enforcement in identifying unregistered or uninsured vehicles, leading to better road safety.
  • Data Insights: Generates valuable data on traffic patterns, peak times, and potential high-risk areas.

Technical Highlights

  • Machine Learning: Employs advanced algorithms for accurate number plate detection and cropping.
  • OCR: Utilizes robust OCR software for reliable character extraction, even in challenging conditions.
  • Real-Time Database Integration: Seamlessly connects to public databases for immediate registration verification.

Inspiration

The inspiration for this project stemmed from the rising incidents of drive-offs, number plate theft, and the increasing need for efficient vehicle monitoring in various settings. We saw an opportunity to leverage existing technology to create a cost-effective solution that could address these challenges.

What it does

Our ANPR system captures images from existing CCTV cameras, identifies and extracts number plate information using machine learning and OCR, and then cross-references this data with a local public database to instantly verify registration status. This provides real-time alerts for unregistered, stolen, or wanted vehicles.

How we built it

We utilized a combination of open-source tools, custom scripts, and cloud-based services:

  • OpenCV: For image processing and manipulation.
  • EasyOCR: For extracting characters from the number plate images.
  • Python: As the primary programming language for scripting and integrating the various components.

Challenges we ran into

  • Data Quality: Gathering and cleaning a diverse dataset of number plate images to train our machine learning model proved to be a time-consuming task.
  • Environmental Factors: Varying lighting conditions, camera angles, and weather conditions made number plate detection and character recognition more complex.
  • Real-Time Processing: Achieving real-time processing speed while maintaining accuracy required optimization of our algorithms and hardware selection.

Accomplishments that we're proud of

  • High Accuracy: Our ANPR system consistently achieves a high rate of accurate number plate detection and character recognition, even in challenging conditions.
  • Cost-Effectiveness: By leveraging existing CCTV infrastructure and open-source software, we developed a solution that is significantly more affordable than traditional ANPR systems.
  • Real-World Impact: Our system has already proven its effectiveness in reducing drive-offs and assisting in the identification of stolen vehicles.

What we learned

  • Importance of Data: The quality and diversity of training data are crucial for the success of machine learning models.
  • Algorithm Optimization: Balancing real-time performance with accuracy requires careful optimization of algorithms and hardware.
  • Collaboration: Working as a team with diverse skill sets was essential to overcoming the challenges and achieving our goals.

What's next for Number Plate Scan

  • Enhanced Analytics: We plan to incorporate more advanced analytics to generate insights on traffic patterns, peak times, and potential high-risk areas.
  • Cloud Integration: Further integration with cloud-based services for scalability and easier deployment.
  • Mobile Alerts: Developing a mobile app to provide real-time alerts to relevant stakeholders (e.g., gas station attendants, law enforcement).
  • Expanding Applications: Exploring the potential use of our ANPR system in other areas such as parking management, toll collection, and traffic monitoring.
Share this project:

Updates