Automatic Licenses Plate Recognition System Garima R. Yadav Dept. of Electronics & Comm. Engineering Marathwada Institute of Technology, Aurangabad (Maharashtra), India yadavgarima08@gmail.com Prof. H.K. Waghmare Dept. of Electronics & Comm. Engineering Marathwada Institute of Technology, Aurangabad (Maharashtra), India hemantkwaghmare@gmail.com Abstract Automatic License Plate Recognition (LPR) is an important research topic in Intelligent Transportation Systems (ITS) involving image processing and pattern recognition, which is a key to implementing the intelligent traffic management. The extracted information can be used at tollbooth for automated recognition of vehicle license plate using a photograph of a vehicle. An automated system could then be implemented to control the payment of fees at parking areas, highways, bridges or tunnels, etc. The proposed model consists of four main parts pre-processing of image, localization of number plate, character segmentation and character recognition. This paper contains new algorithm for locating number plate using Morphological operation, Thresholding operation, Bounding box analysis for number plate extraction, character separation using segmentation. In this paper, our main focus is on license plate detection. Index Terms Pre-processing, Morphological operation, number plate localization, character segmentation. 1. Introduction Automatic License Plate Recognition (ALPR) system is an important technique, used in Intelligent Transportation System. ALPR is an advanced machine vision technology used to identify vehicles by their number plates without direct human intervention. It is an important area of research due to its many applications. The development of Intelligent Transportation System (ITS) provides the data of vehicle numbers which can be used in follow up, analyses and monitoring. ALPR is important in the area of traffic problems, highway toll collection, borders and custom security, premises where high security is needed, like Parliament, Legislative Assembly, and so on. Automatic License Plate Recognition (ALPR) is a form of automatic vehicle identification. It is an image processing technology used to identify vehicles by only their license plates. Since every vehicle carries a unique license plate, no external cards, tags or transmitters need to be recognizable, only license plate. The complexity of smart license number plate recognition work varies throughout the world. For the standard number plate, ALPR system is easier to read and recognize. In India this task becomes much difficult due to variation in plate model and their size. License plate localization part is also very difficult in Indian number plate. So flexible algorithm required for solved this task. Nowadays, vehicles play vital role in transportation. Also the use of vehicles has been increasing because of population growth and human needs in recent years. Therefore, control of vehicles is becoming a big problem and much more difficult to solve. License Plate Recognition systems are used for the purpose of effective control. ALPR process consists of four stages: 1) Preprocessing 2) License plates detection 3) Character segmentation and 4) Character recognition. In first stage, acquired image is enhanced by converting RGB image to gray image, thresolding, median filtering etc. In the second stage, license plate localization which is difficult but most important stage is obtained, based on the features of license plates. Features commonly employed have been derived from the license plate format. The features regarding license plate format include shape, symmetry, height-to width ratio, area in comparison to image size. In third stage, characters are segmented using bounding box analysis. Fourth stage, character recognition can be achieved using back propagation artificial neural network, template matching, nodal analysis etc. In this paper, our main focus is on license plate detection. 2. Proposed system The flow chart of license plate recognition system implementation is shown in the Figure 1. There are various steps in this approach and these are implementation in MATLAB. 1156
Figure.1 Automatic License Plate Recognition System 2.1 Image Acquisition Image Acquisition Image preprocessing License Plate Localization Character Segmentation Character Recognition The first stage of any vision system is the image acquisition stage. The image is taken by low resolution camera or from database according to the requirement of the system. After the image has been obtained, various methods of preprocessing can be applied to the image to perform many different vision tasks. Figure 2. show the captured image of car below. 2.2.1 Converting RGB to Grayscale image It involves conversion of color image into a gray image. It is more convenient and easier to deal with one component (intensity) in gray scale images than three color components (red, green, blue) in color images G(x,y) = 0.3R+0.59G+0.11B The method is based on different color transform. According to the RGB value in the image, get gray value and obtains the gray image. 2.2.2 Median Filtering When images are acquired there is lot of salt and pepper noises present in image. This noise cannot be eliminated in gray image processing. To remove noise from the image median filters are used so, that image becomes free from noise. Noise removal is necessary step in License plate detection because it greatly affects the recognition rate of the system. If f(x, y) represents the dealt image and g(x,y) represents the result image, noise removal using median filtering is shown as: g(x,y)=f(x,y)*h(x,y) Where h(x, y) is filter transfer 2.2.3 Opening & closing of filtered image 2.2 Image Preprocessing Figure 2. Captured image Opening of image is processes of adding pixel to boundary & closing is removing of pixel from boundary. This process combines the nearby white pixel and displays as one region. After subtracting opened image and closed image we get license plate region as connected region with some unwanted region as show figure 3. below. After the acquisition of image, preprocessing of image is done. When an image is acquired, there may be noises, unwanted object etc present in an image. These noises affect the recognition rate greatly. So, these noises should be removed from the images. In image preprocessing various steps are used and these are described below. 1157
2.3 License plate detection 2.2.4 Image binarization Figure 3. Difference image Binarization is a process in which image pixels are distinguished into two kinds of color according to certain criteria. It shows image into two colors black and white so that image is fully clear and without any kind of noise. Figure 4. shows threshold image followed by unwanted region elimination image. Unwanted region is eliminated on the bases of area of connected component in the image. The plate detection is the most important phase of the license plate recognition system. In this, each phase performs a process of segmentation on the threshold image to eliminate the area which does not belong to the license plate region. License plate region are consider based on the features of license plates. Features commonly employed have been derived from the license plate format. The features regarding license plate format include shape, symmetry, height-to width ratio, area in comparison to image size, number of connected component. After all the feature condition is satisfied a bounding box is created on the candidate region and the detected region is segmented. Bounding box is a rectangle box created around the connected object with maximum dimension of object, figure 6. Shows a bounding box mapped image and license plate extraction. Figure 6. Candidate region mapped Figure 4. Binarized image As seen in figure 6. two region are considered as candidate region i.e license plate. After mapping, number of connected component in the mappend region and area is check. If condition of no. of connected component and area is satisfied, candidate region is consider as license plate. And after satisfying all the condition license plate is segmented using deminsion obtained from bounding box. Figure 7. show license plate which extracted from the whole image. Figure 5. Unwanted region eliminated image Figure 7. License plate extraction 1158
2.4 Character segmentation Character segmentation is necessary step for recognition of character. This segmentation technique avoids drawback of vertical projection and connected component analysis of broken character and joint character. characters, different background and foreground colors, etc. The maximum execution time required is 2.060 sec. Below are result for few images. The extracted license plate is dilated first. Dilation is done to avoid segmentation of disjoint character. After dilating extracted license plate, connected component analysis is done i.e each object is labeled and area of each object is checked. Depending upon area, object is eliminated to ovoid dot or any kind of noises figure 8. shows noise free license plate only with character. After obtaining noise free license plate, Figure 10. Result obtain after execution of code 4. Conclusion & future work Figure 8. Noise free license plate and Character mapping bounding box analysis is applied and dimension obtain from bounding box analysis is used for cropping. Figure 9. shows the cropped character. The cropped character are further used as input for character recognition. Figure 9. Character segmentation 3. Experimentation and result The system was tested with a set of images in database. When code is executed in MATLAB a dialog box appears having set of image, from set of images one image is selected as the input image. The images in database is having wide variations in illumination conditions, size of number plate, size of In this paper, we proposed a efficient method for License plate detection. The system has been tested on many images of various lighting conditions & system can be implemented on motorways & highways for automatic toll tax collection. The proposed system works quite well however, there are still areas for improvement. The input image used in this project is taken from average quality & low resolution. This low resolution image can give good result for license plate detection but for character recognition, robustness and good speed high resolution camera should be used. 5. References [1] Rafael C. Gonzalez and Richard E. Woods, Digital Image Processing 3 rd edition, Pearson/Prentice Hall, 2008 (variance) pp. 96 97,(Mathematical Morphology) pp. 628-635, (Segmentation) pp.700-725, (Thresholding) pp.738-742, (Object Recognition - Correlation) pp. 866-872. [2] S. Zhang, M. Zhang, and X. Ye, Car plate character extraction under complicated environment, in Proc. IEEE Int. Conf. Syst. Man Cybern., vol. 5. Oct. 2004, pp. 4722 4726. [3] Z.-X. Chen, C.-Y. Liu, F.-L. Chang, and G.-Y. Wang, Automatic license plate location and recognition based on feature salience, IEEE Trans. Veh. Technol., vol. 58, no. 7, pp. 3781 3785, Sep. 2009 1159
[4] J.-M. Guo and Y.-F. Liu, License plate localization and character segmentation with feedback self-learning and hybrid binarization techniques, IEEE Trans. Veh. Tech., vol. 57, no. 3, pp. 1417 1424, May 2008 [5] X. Fan and G. Fan, Graphical models for joint segmentation and recognition of license plate characters, IEEE Signal Process. Lett.,vol. 16, no. 1, pp. 10 13, Jan. 2009 1160