Crowdsensing the Speed Violation Detection with Privacy Preservation

Abstract

Reckless driving, primarily speeding, poses a significant threat to road safety. While measures such as speed cameras and traffic patrol officers aim to deter speeding, their effectiveness is limited by coverage constraints and their inability to monitor areas beyond their immediate vicinity. To address these challenges, we introduce a crowd-sensing framework for speed violation detection with privacy preservation. This framework enables individuals to identify speeding vehicles while ensuring the privacy of involved parties. The framework comprises three key entities: the Speed Reporter (SR), the Speed Violator (SV), and the Ticket Issuer (TI). By leveraging this approach, the detection coverage of speed violators is extended, providing near real-time monitoring of reckless driving incidents. To correctly identify the speeding vehicles for our framework, we have utilized YOLO (You Only Look Once) to develop a real-time object detection model, for vehicle and license plate detection. The experimental results of our models show the effectiveness and practicality of our proposed framework.