Title :
A robust and highly efficient localization method for irregular license plates
Author :
Ghaffari, B.V. ; Kitajima, T. ; Abdullah, S.S. ; Kouhnavard, M.
Author_Institution :
Malaysia-Japan Int. Inst. of Technol. (MJIIT), Univ. Teknol. Malaysia, Kuala Lumpur, Malaysia
Abstract :
This paper presents a robust method for license plate localization (LPL). Basically, the performance of LPL directly affects the accuracy of License Plate Recognition system. Although many studies had been conducted on LPL systems, many of them only discuss on license plate (LP) from countries that have restricted and distinguished LP types in terms of shapes and fonts. In this paper, a novel LPL system is proposed to address difficulties in recognizing extraordinary LP types with respect to shapes and fonts. The system has been implemented using MATLAB in eight different conditions of illumination in an actual environment to illustrate the efficiency of the proposed method. The total accuracy of this system for all conditions is 96% to 97%.
Keywords :
image recognition; object recognition; traffic engineering computing; LPL systems; MATLAB; license plate recognition system; robust LPL method; robust irregular license plate localization method; Accuracy; Histograms; Image edge detection; Licenses; Noise; Region 10; Shape; Blob analysis; Histogram equalization; Median filter Morphological operators;
Conference_Titel :
Region 10 Symposium, 2014 IEEE
Conference_Location :
Kuala Lumpur
Print_ISBN :
978-1-4799-2028-0
DOI :
10.1109/TENCONSpring.2014.6863109