PyParkOps
May 2024AI/ML
PyParkOps is a computer vision-based parking management system built with Python. It detects vehicles in real time, recognizes Bangla license plates, and automatically calculates parking fees. The system is integrated with a Streamlit dashboard, allowing operators to monitor entries, exits, and earnings live, reducing manual work and errors.
An AI-powered parking system that automates vehicle detection, Bangla license plate recognition, and fee calculation in real time.
Tech Stack
Python
Features
- Real-time vehicle detection using YOLOv8
- Bangla license plate recognition with EasyOCR
- Automatic entry/exit tracking and fee calculation
- Live dashboard with camera feed and reports
Challenges
- Bangla text recognition was not always accurate
- Issues with image quality affecting detection
- Fixing logic for correct fee calculation
- Handling real-time processing smoothly
Learnings
- Learned how computer vision pipelines work end-to-end
- Gained experience with YOLO and OCR integration
- Improved handling of real-time data and image processing
- Understood practical challenges of ML accuracy
Highlights
- Reduced manual work in parking management
- Combined CV, OCR, and dashboard into one system