Lot Vision

Smart Parking, Smarter Security: AI-Driven Access Control

Comment

Inspiration

Our project was inspired by the need to streamline the cumbersome and costly manual parking enforcement processes common in various parking lots. Traditional patrol methods are not only inefficient but also environmentally unsustainable. We aimed to leverage advanced technology to improve parking fairness, reduce operational costs, and enhance overall security and efficiency in any parking facility.

What it does

The AI-Powered Smart Parking Enforcement system automatically detects and reads license plates at entry and exit points of parking lots. It cross-checks captured plates against an authorized vehicle database in real-time, instantly logging and reporting unauthorized vehicles. Parking enforcement personnel can remotely address violations, eliminating the need for continuous physical patrols.

How we built it

We developed a compact, affordable hardware setup using Raspberry Pi 5 integrated with a high-resolution camera. This setup captures real-time license plate images. The system utilizes OpenCV and PaddleOCR for accurate, high-performance detection and recognition of license plates from video streams. The backend is powered by FastAPI and SQLite, structured to manage data efficiently across tables for registered vehicles, active parking lots, occupancy status, and violation logs. For frontend development, we created an interactive, user-friendly dashboard using React and TailwindCSS, seamlessly integrated with backend APIs. Additionally, we employed Gemini 2.0 Flash to analyze vehicle logs, providing insightful alerts about potential violations or unusual patterns.

Challenges we ran into

One significant challenge was ensuring accurate and rapid license plate recognition under varying environmental conditions. We extensively tuned PaddleOCR and OpenCV algorithms to maintain reliability and performance. Integrating multiple technology stacks, from hardware capture to frontend visualization—required precise coordination, thorough testing, and debugging.

Accomplishments that we're proud of

We successfully developed a comprehensive, robust system capable of significantly reducing patrol costs and environmental impacts through automation. We created a reliable, scalable platform that performs efficiently under real-world conditions, proving its feasibility for practical implementation.

What we learned

Throughout this project, we gained deep insights into hardware-software integration, particularly with Raspberry Pi setups. We refined our skills in OCR technology, advanced machine learning techniques, real-time image processing, and improved our capabilities in full-stack development, API integration, and user interface design.

What's next for Lot Vision

We plan to enhance our system by:

  • Implementing an SMS/email notification system.
  • Expanding the admin dashboard functionality for improved personnel experience.
  • Integrating with existing consumer ID systems.
  • Developing a live incident heatmap to visualize and anticipate parking enforcement needs.

Built With

Share this project:

Updates