Number Plate Recognition System For Indian Vehicles

Number Plate Recognition System For Indian Vehicles

Rishabh Agrawal Computer Engineering, University Of Pune

Saurabh Sharma Computer Engineering, University Of Pune

Shayani Gupta Computer Engineering, University Of Pune

Abstract

Vehicle number plate recognition (NPR) has been intensively studied in many countries. Due to the different number plates being used, the requirements of a number plate recognition system are different for each country. Project implements a number plate localization and recognition system for vehicles in India. The system is based on digital images and can be easily applied to commercial car park systems for the use of documenting access of parking services, secure usage of parking houses and also to prevent car theft issues. System recognizes the number plates written in standard format. Different algorithms are implemented for different functions of the system. Adaptive Otsus Method is used to binarise the grayscale image. Cannys Edge Detection is used to detect the edges. Feature Extraction is used to localize the number plate from the image. Back propagation neural network will be used for recognition of the characters. These various algorithms work efficiently for the Indian conditions hence overall system efficiency increases.

1. Introduction

   Number Plate Recognition (NPR) is probably the next growth product to take off in the whole world. NPR is a mass surveillance method which works on images to read vehicle registration plates. They can use existing closed-circuit television or ones specifically designed for the task. They are used by various police forces and as a method of electronic toll collection on pay-per-use roads and cataloguing the movements of traffic or individuals. NPR can be used to store the images captured by the cameras as well as the text from the license plate. They are meant to detect or prevent any sort of activity from unauthorized

   people. They are necessarily meant for improving the security.

   In tradition, a general recognition method of supervisory vehicle apparatus usually uses human to recognize the license number monitored by a camera. Basically, using human to recognize vehicle license number in this way may solve the problems. But using people to do the recognition takes time and needs many workers to do this work. Hence, NPR is developed for the vehicle identification to help a human operator and improve the service or work quality. Some applications of an NPR system are, automated high-way toll collection systems, automation of petrol stations, journey time monitoring.

   The task of recognizing number plate in Indian conditions becomes a bit difficult as many Indian vehicles do not follow the standard format of the number plates. Indian number plates can have single row or double row. It can also have few types of variations.

   The existence of a multitude of variations across number plates adds a considerable amount of complexity to the process of reading them. In order to offer a reliable solution in Indian conditions where number plates are far from standardized, it has two modules that comprise NPR. In many cases, the number plate/license plate of a vehicle is not affixed in the prescribed location. The Number Plate Extraction tool automatically analyzes an image of the vehicle and locates the position of the license plate and extracts the same and stores it. Once the Number Plate Extraction extracts the license plate from an image of the vehicle, this is passed to the Number Plate Reading module that reads the numbers and characters on the license plate and passes it onto a computer. This module can be integrated with a database for further functionalities.

2. Problem Formulation

   1. Problem Definition Given the input image of vehicle number plate, system should extract that number from the image and should search the database for that recognized number plate. The number plate recognition should be robust i.e. It should recognize the number plates even in the low light or shadow like conditions.

   2. Problem Output – The main objective of NPR system is to recognize the number plates reliably and with greater efficiency without any human intervention. The functionality of the system provides the ability to track a vehicle in any state, to find a stolen vehicle ,or for managing the parking in the malls etc.

      Real time inputs are taken from the front camera. The images are processed and the number plate will be detected. The detected number plate is then be used to search in the database to see if it exists or not.

3. Mathematical Modelling

   We now provide a model of the system in terms of Set Theory domain.

   1. Let us consider S as a system for Number Plate Recognition System FOR Indian Vehicles including the set of inputs, set of outputs, set of function, set of limitation, set of condition.

   2. S= {

   1. Identify the inputs

      S= {I..}

      Input set:

      I={R}

      • R is the set of image given as the input to the NPR system.

        Ri= {i |i images retrieved from database}

        = {i | i=1}

   2. Identify the outputs S= {I, O}

      Output set:

      O= {RN, RS, PI}

      • RN is the set of recognized number from the number plate.

RNi= {i | i is the recognized number }

• RS is the set of records associated with the number plate

  RSi= {i | i is the number of records}

  = {i | i=1}

• PI is the set of image after preprocessing of the number plate.

  PIi = {i | i is the preprocessed image}

  1. Identify the functions

     S = {I, O, F }

     Function set:

     F = {M1, M2, M3}

• M1 is the set of Functions used for preprocessing of the input image

  M1= {E, B, L, S, SG}

  • E ( ) enhances the features of the image captured by the camera.

  • B( ) is a function that is used to binarise the enhanced image

  • L ( ) is a function that is used to locate the number plate in the image.

  • S ( ) is a function that is used to sharpen the image.

  • SG ( ) is a function that segments the characters of the localized number plate.

    M1 (Ri) =PIi

    M1i = {i |i the preprocessed image} Where M1 preprocesses on the image Ri and gives the output as preprocessed image Pi.

    M1

    M1

    Ri PIi

• M2 is the set of functions used for recognition of characters from preprocessed image.

  M2= {Rec}

  • Rec recognizes the characters of the number plate.

M2 (PIi) =RNi

M2i={i | i are the recognized characters } Where M2 recognizes characters from the image Pi and gives the output as Ri

M2

M2

.

PIi RNi

• M3 is the set of functions used for searching and displaying records from the database.

  M3= {SE, D}

  • SE searches the recognized number plate in the database.

  • D displays the details corresponding to the recognized number plate

M3 (RNi) =RSi

M3i= {i | i are the recognized characters}

Where M3 searches and displays the details of the recognized number plate the image Ri and gives the output as Rs .

M3

M3

RNi RSi

1. Identify the Limitations S = {I, O, F, P}

Limitation Set: P= {Fz, Br}

• Fz states that the number plate to be recognized should not be fancy.

• Br states tha the number plate to be recognized should not be broken.

1. Framework

   The basic framework for Number Plate Recognition consists of 3 main phases:

   Phase 1(Receive Input image) Receives the image from the camera and sends them to the next phase.

   Phase 2(Process image) Preprocesses the image to enhance its features and to binarise it, number plate localization to detect the location of number plate in the image, character segmentation, character recognition through back propagation neural network .

   Phase 3(Display Information) Receives the recognized Number and it is searched in the database to get the

   information about that particular vehicle, then this information is displayed.

2. Image Processing

   The following functions shall be used to process the image.

   1. PreprocesRsiNnig

      RSi

      RSi

      The RGB image is converted into a greyscale image for easy analysis as it consists of only two colour channels. Filtering is then implemented for the effective removal of speckle noise, salt and noise pepper noise (impulsive noise)[1].It enhances the features of the image captured by the camera using the function E().

      Figure 1: Grayscale Image

   2. Binarization

      After the preprocessing of the image binarization of the image has to be done so that further image processing becomes easier. Binarization of the image is implemented using adaptive Otsus method[6] which considers the shadow and light intensity in the image. This binarised image will be used for the purpose of localizing the number plate. B () is a function that is used to binarise the localized image.

      Figure 2: Binarized Image

   3. Plate Region Extraction (localization)

For the purpose of number plate localisation in the image Feature Extraction has been used. We have also made use of Cannys method for edge detection as it has an accuracy of 98% which is far more than the other edge detection methods. We have made use of the features of a number plate[4]. A number of candidate rectangles are found out. The candidate rectangles are found out by finding out those regions having 11 continuous white vertical pixels and 18 continuous white horizontal pixels.

Figure 3: Localised Image

1. Segmentation

   Characters present in the number plate have to be segmented so that each character can be recognized separately. This segmentation will be implemented by the method of projections[5] .Horizontal and vertical projections are used for segmenting each character[4]. These segmented characters will then treat as the input for the neural network. SG() is a function that segments the characters of the localized image.

   Figure 4: Segmented Image

2. Recognition

Back propagation neural network is used for the recognition of the individual characters[3]. Neural network will consist of various nodes with different classifiers. Neural network is trained under supervised training where a teacher is needed and unsupervised training where no such person is needed and the network tries to learn it.M2 is a set of functions that is used to recognize the characters in the number plate.

Result

The result is the successful recognition of the number plate of a vehicle which is searched in the database for detailed information.

Conclusion

Thus we have presented a paper for number plate recognition using different algorithms. We have also described the Mathematical Model for the same.

Acknowledgement

This project cannot be considered complete without mentioning Prof. P.M Kamde. We wish to express true sense of gratitude towards his valuable contribution. We are grateful to him for his constant encouragement and guidance in the fulfillment of the project activity.

References

1. Shyang-Lih Chang, Li-Shien Chen, Yun-Chung Chung, and Sei-Wan Chen,Automatic License Plate Recognition , IEEE Transactions on Intelligent Transportation Systems, VOL. 5, No. 1, March 2004.

2. M. M. Shidore and S. P. Narote,Number Plate Recognition for Indian Vehicles, IJCSNS International Journal of Computer Science and Network Security, VOL.11 No.2, Feb. 2011.

3. Stuti Asthana, Niresh Sharma, Rajdeep Singh, Vehicle number plate recognition using multiple layer back propagation neural networks, International Journal of Computer Technology and Electronics Engineering (IJCTEE) Volume 1, Issue 1.

4. Kumar Parasuraman, Member, IEEE and P.Vasantha Kumar, An Efficient Method for Indian Vehicle License Plate Extraction and Character Segmentation, 2010 IEEE International Conference on Computational Intelligence and Computing Research.

5. Mei-Sen Pan, Jun-Biao Yan and Zheng-Hong Xiao, Vehicle License Plate Character Segmentation, International Journal of Automation and Computing 05(4), October 2008, 425-432

6. Yudong ZHANG and Lenan WU, Fast Document Image Binarization Based on an Improved Adaptive Otsus Method and Destination Word Accumulation, Journal of Computational Information Systems 7: 6 (2011) 1886-1892.