Wrong-way driving is one of the primary causes of traffic jams and road accidents. The accurate detection of wrong routes in real-time, combined with prompt security alerts, can significantly enhance safety measures. Our research introduces a solution leveraging You Only Look Once version 8 (YOLOv8) and centroid tracking algorithm, along with Optical Character Recognition (OCR) techniques, for real-time identification of wrong-way vehicles and generation of security alerts.
This project aims to enhance road safety by detecting vehicles taking wrong routes. It utilizes YOLO v8 for real-time vehicle detection and Optical Character Recognition (OCR) to read number plates. When a violation is detected, alerts are sent to the nearby traffic control room and the vehicle owner.
- Real-Time Vehicle Detection: Utilizes YOLO v8 for accurate and efficient vehicle detection.
- Number Plate Recognition: Uses OCR to extract the number plate information from detected vehicles.
- Alert System: Sends alerts to nearby traffic control rooms and the vehicle owner in case of a violation.
- Python 3.7 or higher
- pip
- CUDA (if using GPU)
- cd wrong-route-vehicle-detection-with-Security-Alert
pip install -r requirements.txt YOLO v8 Setup
Follow the instructions to set up YOLO v8 from the Official Repository.
wrong-route-vehicle-detection/
- Video files/ - All datasets for testing
- yolo/ - Pre-trained YOLO v8 models
- ocr/ - OCR-related scripts and models
- requirements.txt - List of dependencies
- README.md - Project documentation
- YOLO_VID_SHORT - Demo video to show yolo application
- ocr - Demo image to show ocr application
- Application of system in rural areas
- Application of system in too busy area with too much traffic
- Application of gaze estimation and emotion analysis to predict accidents and cause of wrong route movements
-
YOLO v8
-
Tesseract OCR