Automatic Recognition of Nigerian Licensed Vehicle Plate Number using Matlab
Lasisi, H.1, Abdulmajeed, M.K.2, Obiyemi, O.O3, Olasege, I.O4.Oladepo,O.5
1,3,4, 5Dept. of Electrical & Electronics Engineering, Osun State University, Osogbo,
2Dept. of Systems Engineering, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia.
Email: lashammed@yahoo.com1,kbmajeed12,obiseye.obiyemi@uniosun.edu.ng3, olasegeismail@gmail.com4, ooladepo@yahoo.com 5
Corresponding Author: Oladepo, O.
ABSTRACT
This paper focuses on the recognition of Nigerian licensed plate numbers using template matching. There are three stages in this work; in the first stage, the image of a licensed plate is loaded in Matlab, converted to grayscale and binarized using ‘Otsu’. It is then filtered, filled and resized. In the Second stage, segmentation was done by identifying blobs, cropping them out using machine vision border-box feature and then resized to pixels – same as the templates. Finally, Template matching was done using the ‘sum of absolute differences’ technique. The recognized licensed plate characters were displayed in a command window. This system promises to be of high accuracy provided the templates are properly constructed. Finally, the system was tested on five different plate numbers with character recognition accuracy ranging from 75% to 93.75% while a modified algorithm gave accuracies of 93.75% to 100%.
Keywords:Licensed plate recognition, Image Processing, Thresholding, Segmentation, Template Matching, Sum-of-Absolute-Differences.
