Implementation of Convolutional Neural Network-based Models for Optical Character Recognition of Nigerian License Plates

Implementation of Convolutional Neural Network-based Models for Optical Character Recognition of Nigerian License Plates

Poster 215

The last stage of the license plate recognition system is optical character recognition, and this has become a subject of concern. A lot of research has been going on in recent times because of the ability to convert from human-readable to machinable form without changing, noise variations, and several other factors, which largely depend on the quality of the input documents. License Plate comes in various sizes, colors, designs, and shapes to be attached to the front or rear of a vehicle and depending on the user or nature of the work being served to determine what type of plate number goes for what type of vehicle. In Nigeria, there are about eight (8) different types of License Plates in Nigeria,, ranging from (1)customized/fancy plate numbers,(2) commercial Plate numbers,(3) Private plate numbers, (4)Government official plate numbers, (5) Armed Forces plates number, (6) Temporary plate number and (8) each of the plates number is differentiated by either the color of the lettering, design on the plate or the background of the plates. This Research intends to be able to Recognize each of these plates numbers using the License Plate Recognition system, this is one of the major components of the Intelligent Transport System.