Abstract— A car has a unique label called a license plate or a vehicle registration plate. A vehicle registration plate (VRP) in Indonesia consists of a series of alphanumeric. The aim of this research is to recognize the VRP from a frontal image of a car. Image processing techniques are done to identify the VRP, by converting to grayscale image, transforming into binary image by using adaptive threshold, applying region labeling, filtering and others image processing techniques. Features of a character from VRP are taken by applying the Discrete Cosine Transform (DCT) then the character is recognized by Radial Basis Function network (RBF). Some experiments are conducted to determine the number of DCT is needed for each character and to know how well the system to recognize the VRP. System uses 63 DCTs to each character and recognizes the VRP with percentage is about 96%.
Index Terms— Discrete cosine transform, radial basis function network, vehicle registration plate.
P. Yugopuspito, S. Lukas, Sutrisno, and D. Krisnadi are with Informatics Department, Faculty of Computer Science, Universitas Pelita Harapan, UPH Tower, Lippo Village Karawaci, Tangerang 15811, Indonesia (e-mail: pujianto.yugopuspito@uph.edu, samuel.lukas@uph.edu, sutrisno.fik@uph.edu, dion.krisnadi@uph.edu).
[PDF]
Cite: Pujianto Yugopuspito, Samuel Lukas, Sutrisno, and Dion Krisnadi, " Recognition of Indonesian Vehicle Registration Plate by Discrete Cosine Transform and Radial Basis Function Network," International Journal of Information and Electronics Engineering vol. 5, no. 4, pp. 275-279, 2015.