In this paper, we propose the new method to enforce the tailgating violations that occurs at the intersection. The tailgating enforcement system is composed of the wide-view camera for tracking a vehicle and the narrow-view camera for recognizing a license plate. The images from the wide-view camera are sequentially stored in the stack in order to monitor the ROI (Region of Interest) at the intersection area. The narrow-view camera recognized the license plate. And the image coordinate of the license plate from the narrow-view camera was converted into the image coordinate from the synchronized wide-view camera. The feature points of the vehicle were extracted the image from wide-view camera using the converted image coordinate of the license plate from the narrow-camera. The images from the wide-view camera are searched in reverse order, and the vehicle trajectories are tracked. As a result, the success rate of the back-tracking is 98% at day and 93% at night. And the success rate of the sequential tracking is 90% at day and 76% at night. The back-tracking is the better results than the sequential tracking. Therefore the back-tracking is the appropriate to enforce the tailgating violation at the intersection.
CITATION STYLE
Choi, S. H., Ahn, J. P., Rheu, J. H., & Kim, Y. M. (2015). Tailgating enforcement based on back-tracking in intersection. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9490, pp. 630–637). Springer Verlag. https://doi.org/10.1007/978-3-319-26535-3_72
Mendeley helps you to discover research relevant for your work.