Traffic Rules Violation Detection using Dashcam Footage

Traffic Rules Violation Detection using Dashcam Footage

Anvesh J1, Dr. Ramesh B2, Ashrith H3, Bhuvan M D4, Gagan H C5

1Department of Computer Science and Engineering, Malnad College of Engineering, Hassan

2Professor, Department of Computer Science and Engineering, Malnad College of Engineering, Hassan

3Department of Computer Science and Engineering, Malnad College of Engineering, Hassan

4Department of Computer Science and Engineering, Malnad College of Engineering, Hassan

---------------------------------------------------------------------***---------------------------------------------------------------------

Abstract - Traffic Rules Violation Detection using Dashcam Video Footage offers a novel strategy for improving traffic management and road safety through automated violation identification and reporting. The project's goal is to create a video processing system that can use dashcam data to identify and classify frequent traffic infractions. The system can recognize traffic infractions including speeding, lane changes without signaling, helmet recognition, and instances of triple riding on two-wheelers with high accuracy thanks to its integration of YOLO, OpenCV and RESNET50 classifier. The project also includes functionality for user account management and the implementation of a secure user authentication system. Users can examine and validate complete violation reports generated by the system prior to submission. The outcomes show a low false positive rate, rapid real-time processing, and good detection accuracy. Law enforcement organizations have shown. This initiative is an example of how important technical innovation is to solving societal issues and advancing public safety.

Key Words:  YOLO (You Only Look Once), OpenCV, ResNet50 (Residual Neural Network), Dashcam Video Footage.