DocumentCode
143541
Title
Saudi license plate localization using hybrid method
Author
Almustafa, Khaled ; Zantout, Rached
Author_Institution
Coll. of Eng., Prince Sultan Univ., Riyadh, Saudi Arabia
fYear
2014
fDate
13-16 April 2014
Firstpage
236
Lastpage
241
Abstract
In this paper, localizing Saudi license plates in images using the “character recognition” method is described. In this method, all possible objects from the license plate image are filtered to keep only objects which may belong to a Saudi license plate. Filtering is done by checking each object to see whether it is a possible letter or number (in Arabic or English) that could be part of a Saudi license plate. If there exist both Arabic and English patterns that agree, the license plate is recognized. The method correctly localized and recognized Saudi license plates 83.9% of the times. The cases where the method did not work were for noisy images of license plates.
Keywords
character recognition; object tracking; traffic engineering computing; Arabic patterns; English patterns; Saudi license plate localization; character recognition; image filtering; license plate image; Character recognition; Conferences; Educational institutions; Feature extraction; Image recognition; Licenses; Object recognition; Topological relationship; character recognition; license plates localization;
fLanguage
English
Publisher
ieee
Conference_Titel
Mediterranean Electrotechnical Conference (MELECON), 2014 17th IEEE
Conference_Location
Beirut
Type
conf
DOI
10.1109/MELCON.2014.6820538
Filename
6820538
Link To Document