期刊名称：TELKOMNIKA (Telecommunication Computing Electronics and Control)
印刷版ISSN：2302-9293
出版年度：2018
卷号：17
期号：2
页码：620-627
DOI：10.12928/telkomnika.v17i2.9017
语种：English
出版社：Universitas Ahmad Dahlan
其他摘要：The main principle of license plate recognition is to recognize the characters in the license plate which indicates the identity of the vehicle. This research will provide a system which can be implemented to the automatic payment in highway. Indonesian license plate consists of two parts, every of which has certain characters. These characters may become problem in the recognition process. Another problem is on the type of the license plate since Indonesia applies different color for every type of vehicle. In this research, different approaches are employed in the recognition of license plate; that is using character area as the feature value, also known as feature area, and K-Nearest Neighbor (KNN) as classification method. In addition, another method that has been used in our previous research is also employed to detect the character of license plate. The result shows very significant accuracy of 99.44%. In the process of recognition, scenario 1 gives the best accuracy at the K-1 value; that is 68.57% on the license plate and 92.72% on the characters of license plate. In the scenario 2 was obtained the license plate accuracy of 52% and license plate character accuracy of 89.36% with K-5. The system ran in a relatively short computational time.
关键词：area feature extraction;character recognition of license plate;K-Nearest Neighbor;license plate recognition
