Abstract-Automatic license plate recognition systems (LPR) can help reduce the number of traffic violations and make our streets safer. In this project we developed an automatic system that locates Saudi license plates in a captured image regardless of the time of day or license plate scale. The proposed system can tolerate slight tilting of the license plate. The localization process is fairly complex due to the highly varying nature of the background. Good results were obtained using the localization stage. A second part of the system was developed to segment and recognize the characters in the located license plate. Khalil [5] were all on the previous Saudi license plates series.
Index Terms-IntelligentMaglad [6], [7] proposed a system that uses radial basis function neural networks for the detection and recognition of plates. Khalid Almustafa [8]- [10] proposed an algorithm that uses pixel density to recognize characters in license plates. Shaaban [11] proposed a system where a neural network classifier is used to recognize the characters based on edge moment invariants and principal component analysis features of wavelet coefficient matrix.Alginahi [12] proposed a system that recognized both old and new Saudi license plates. The license plates are detected by searching the vertical lines then the width to length ratio is used to locate the plate. Character segmentation is then performed before both a distance classifier and a Neural Network classifier are used to recognize characters.Deriche [13] proposed a hybrid system of RFID and license plate recognition. The license plate detection is based on morphological filtering and the character recognition is based on neural networks. Mansour [14] proposed a system based on chain code and tested it on both Egyptian and Saudi license plates. decade. Figure (1 1950, 1962, 1972, 1978, 1996 and 2007. III. PROPOSED SYSTEM
II. SAUDI LICENSE PLATES
A. LocalizationLocating the license plate in an image is the first step in any license plate recognition system. Variation in license plate location in different cars, variation in location of cars in images and variation in lighting during day/night makes this step challenging.We started looking into locating the plates by searching for black lines that borders the license plates thus we have to look for a horizontal line and a vertical line, however we International Journal of Computer and Electrical Engineering, Vol. 5, No. 1, February 2013 found that in many cars that border is hidden under a frame that is commonly used to hold the plate.We have noticed that particular to Saudi license plates, there is a black cross that centers all license plates, this cross is visible in all plates and is not hidden by frames since is it is in the center of the license plates.We apply an edge detector to find the horizontal and vertical maps; we then use the average value of pixels in each map as a threshold to convert it into a binary image and use a median filter to remove noise and find the binary addition of the maps. To h...