<![CDATA[One of the causes of accidents is the lack of vigilance among drivers and violations of vehicle speed exceeding the maximum limit. To mitigate these violations, traffic supervision is necessary, especially in accident-prone areas. This research introduces a video-based vehicle speed and license plate detection system developed using YOLOv5-DeepSORT and HyperLPR to address this issue. The system employs YOLOv5 and DeepSORT to detect and track vehicle movements, thereby obtaining the displacement of the vehicle, which serves as a reference for speed detection. HyperLPR, on the other hand, is utilized for license plate recognition. The research adopts an experimental methodology, involving video recording on the Cipali toll road section, which serves as input for the vehicle speed and license plate detection program. The evaluation of vehicle object detection using YOLOv5 yields a precision metric score of 100%. Moreover, the testing of vehicle speed detection reveals an average error percentage of 7.6% compared to the actual values. In terms of license plate detection, an overall character accuracy rate of 91.82% is achieved.Based on these findings, it can be concluded that the developed vehicle speed and license plate detection system exhibit excellent accuracy and could be considered for implementation, taking into account predefined implementation criteria.Keyword: Vehicle speed, vehicle license plate, YOLOv5, DeepSORT, HyperLPR.]]>
Copyrights © 2023
