Iranian License Plate Recognition based on global feature extraction

Neda Nikmehr (P50199)

Please use this identifier to cite or link to this item: https://ptsldigital.ukm.my/jspui/handle/123456789/476118

Full metadata record

DC Field	Value	Language
dc.contributor.advisor	Md .Jan. Nordin, Prof. Dr.
dc.contributor.author	Neda Nikmehr (P50199)
dc.date.accessioned	2023-10-06T09:13:46Z	-
dc.date.available	2023-10-06T09:13:46Z	-
dc.date.issued	2012-11-07
dc.identifier.other	ukmvital:73414
dc.identifier.uri	https://ptsldigital.ukm.my/jspui/handle/123456789/476118	-
dc.description	License Plate Recognition (LPR) techniques are a significant part of Intelligent Transportation System (ITS). This system was designed to make human work easier, and can reduce the uses of human resource and cost, because of its potential in applications, such as parking entrance, exit management, highway surveillance, electronic toll collection, security control for forbidden areas, border control and law enforcement. The development of automatic recognition of license plates leads to greater efficiency for vehicle control. The algorithms which are used in Iranian license plate recognition are so complex and time consuming because of numbering and the complexity of the Iranian characters contained in its alphabet. Despite having high accuracy their computational cost is high. So in this research, an Iranian License Plate Recognition (LPR) is developed based on Global Feature Extraction that is more effective and efficient on computational costs, time and it has a high accuracy rate. Four image classification techniques will be discussed related to Character Recognition (CR). The experiment is done by common image processing analyze tools. This system is focused on the Iranian type of vehicle plate. The LPR system performance has been tested on 30 true color vehicle images which are captured by Gatso speed camera on the Tehran Ghom Highway in Iran. The research consists of four parts. Firstly, License Plate Localization based on morphology operation. Secondly, character segmentation is done by connected component labelling. Then, we proposed Edge Direction Matrixes (EDMS), Gray Level Co-occurrence Matrix (GLCM), and GLCM with feature selection and Combination of GLCM by feature selection with EDMS as Feature Extraction methods. We applied these Feature Extraction methods on 151 Iranian character images which involve alphabet characters and numbers. Finally, CR is developed using five different classifiers to recognize the characters of Iranian car license plates. Bayes Network, Multilayer Network, AdaboostM1, Decision Tree and Support Vector Machine (SVM) are used as classifiers. High accuracy rate of correctly classified and time taken have been evaluated after the test performance of CR with four classifiers. EDMS with Support Vector Machine (SVM) classifier has a slower time taken in comparison with other methods and it has a high accuracy rate approximately 90 % for Iranian license plate recognition. Thus, the experimental results showed the specific performance of EDMS feature extraction method in comparison with other methods with Support Vector Machine (SVM) classifier.,Master/Sarjana
dc.language.iso	eng
dc.publisher	UKM, Bangi
dc.relation	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat
dc.rights	UKM
dc.subject	Iranian
dc.subject	License Plate Recognition
dc.subject	Global feature extraction
dc.subject	Intelligent transportation systems -- Iran
dc.title	Iranian License Plate Recognition based on global feature extraction
dc.type	theses
dc.format.pages	87
dc.identifier.callno	TE228.3.N536 2012 3
dc.identifier.barcode	000340
Appears in Collections:	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Files in This Item:

File	Description	Size	Format
ukmvital_73414+Source01+Source010.PDF Restricted Access		1.88 MB	Adobe PDF	View/Open

Show simple item record Recommend this item