Design and build motorbike parking management system based on license plate recognition

Abstract

As the number of vehicles has been quickly increased, automatic parking management has been highly demanded and studied. This project proposes a complete motorbike parking management system (PMS) solving main current issues of manual slow entry-exit checking, missing information storage and statistic data record. The proposed system consists of IP camera, license plate recognition module, RFID card reader and web-based management application. The major steps are detecting the license plate through WPOD-NET Model, reducing the noise by Gaussian Filter, using contour to separate characters from license plate image and then training the Support Vector Machine algorithm for recognition. Recognized license data from the recognition module is then compared with stored radio-frequency identification (RFID) information for entry or exit permission. A web-based management application has been built using Laravel framework for managing individual database consisting of license plate information, driver’s face image, parking fee, entry and exit times, and presenting all related information on display interface. The administrator will access the local server to assign permissions to each user. The recognition results show the high recognition rate at about 97% under different operation environments. The recognition process and operation time is less than 1 second. Further analysis on other criteria of the parking management system has been presented in this study