ABSTRACT
The aim of this project is to design license plate recognition (LPR) system
where the recognition of the plate number will be performed by using personal
camera and personal computer. The recognized plate numbers will be stored in the
text file. This project differs from the existing products where the off-the-shelf
components are used to optimize the cost. Five steps of image processes are adapted
to produce reliable results which are image acquisition, image pre-processing, plate
localization, plate segmentation and plate recognition. In order to ensure the accuracy
of the characters detected to match with the characters stored in database, the
Adaptive Thresholding algorithm has to be designed properly. Based on this
approach, a LPR system is successfully developed with accuracy of 91%