International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 4, Issue 5, May 2015
ISSN 2319 - 4847
Vehicle Plate Extraction and Recognition using
Hopfield Neural Network and Comparison with
DWT, Correlation and NN Algorithms
1
Er. Gurjinder Pal Singh , Er. Navneet bawa2
1
Department of CSE, DIET, Kharar ( Pb.), India
2
Department of CSE, ACET, Amritsar ( Pb.), India
ABSTRACT
Advances in the technology in all aspects of modern world leads to the development of new methods for information security
and monitoring system. Surveillance system is used for security in various fields like home security, traffic monitoring and toll
collection. Automation of various systems leads to better monitoring of the system. Traffic monitoring and security system uses
vehicle number plate identification for their owners who disobey the traffic rules, stolen vehicles and speed monitoring. This
paper presents a novel approach for one search technique to identify vehicle no plate using Hopfield neural network The NN
system is trained using all the characters in different style and sizes so that the system is made independent of size, rotation and
location. Hopfield Neural Network is based on image pixel pattern. The proposed algorithm is compared with correlation based
methodology and artificial neural network.
Keywords:
VNPR, Segmentation,
K-means Clustering, Hopfield Neural Network.
1.INTRODUCTION
Security on roads with respect to vehicle tracking system has become an important issue in today’s high traffic
population. Vehicle tracking is one of the important aspect in vehicle security. High security number plate system has
been incorporated in several countries so that the vehicle could be traced by tracking or reading the integrated chip
installed on number plate. However, the text on number plate has been the most convenient way when doing the same
manually. The presented work that ease the manual operation by just using a smart camera equipped with the
developed algorithm enables the traffic personnel to track the vehicle under scanner. Traditional monitoring and
security systems are slower, prone to errors and require human power 24x7 time at the location. These systems have
many limitations like attention, delayed information and time consuming. Vehicle number pate recognition (VNPR) is
an advanced methodology in which vehicles are identify from their owner by using number plates. In India vehicle
number plate is composition of ten characters. Indian number plates have four parts: State, State Code, District
information and Registration Number. State represented by two alphabet characters, state code represented by two
digits ,district information represented by two alphabets and registration number represented by last four digits This
system can be used in various purposes: monitoring administration, urban road monitoring ,border crossing control,
indoor parking or underground parking.
2.RELATED WORK
Singh Gurjinder pal et al . (2015) described Number plate identification using image processing methodology is used
for extracting and identify vehicle by reading through number plate. Identification is an essential area in the
development of intelligent traffic management systems and surveillance. [1].
Singh Gurjinder pal et. al (2015) Vehicle Number Plate Recognition system has gained wide popularly with the
continuous increase in the number of vehicle related offences. Its research is becoming challenging and interesting day
by day. VNPR is designed to help in recognition of number plates of vehicles. Number plate recognition is the term
used to unique identify road vehicles without human intervention.[2]
Nandan More et.al(2015) described License Plate Identification Using Artificial Neural Network and Wavelet
Transformed Feature Selection is proposed, which is based on wavelet transform function and comparing the proposed
methodology with correlation based method for the detection of number plate. Experimental result shows that the
proposed technology has the better performance in comparison with the correlation based methodology. The matching
process for number plate recognition is modified with the help of multi-class RBF neural network optimization[3]
Jitendra Sharma et.al (2014) described a new methodology for ‘License Plate Recognition’ based on wavelet
transform function. This proposed methodology compare with Correlation based method for detection of number plate.
[4].
Volume 4, Issue 5, May 2015
Page 374
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 4, Issue 5, May 2015
ISSN 2319 - 4847
Ayman Rabee et. al. (2014) proposed a highly reliable license plate detection and recognition approach using
mathematical morphology and support vector machines (SVM). [5]
Cosmo H.Munuo1 et. al. (2014) described this paper is a review of NPR preliminary stages; it explains number
plate localization, sizing and orientations as well as normalizations sections of the Number Plates Detection
and Recognition.[6]
Mahmoud Ibrahim et.al (2013) defined ALPR used either a color, black and white, or infrared camera to take
images. The quality of the acquired images isa major factor in the success of the ALPR. ALPR as a real-life application
has to quickly and successfully process license plates under different environmental conditions, such as
indoors,outdoors, day or night time. [7]
Firas Mahmood Khaleel et.al (2013) described Speeded Up Robust Features (SURF) and Bag-of Words (BoW)
feature descriptors are combined and clustered by using K-means clustering to form a novel way of localizing the
license plate’s region in an image. [8]
Guan Xuezhong et. al(2013) described this article mainly researches on license plate automatic recognition technology
which is used in parking management system. LPRS could replace the traditional methods like swiping card or
artificial recognition in license plate recognition, the process is divided into four stages: image pre-processing, license
plate localization, character segment and character recognition, and to set up a software system to verify at the same
time .[9]
Bing-Fei Wu et.al(2013) The proposed DLPR system adopts Histogram of Oriented Gradient (HOG) and Support
Vector Machine (SVM)-for classification-based solution. The proposed system implements a two-level LPR, including
LP localization and character recognition[10].
Jian Yang et.al (2013) developed a system which included the image acquisition, license plate location, character
segmentation and character recognition. To improve the accuracy of the license plate location, we improved the effect
of the binary image positioning method.[11]
Yingjun Wu et.al (2013) A novel LPL algorithm is proposed in this paper to jointly consider color and texture
distribution features of license plate area (LPA) in an image. The method is divided into two steps of detecting
vehicles through background subtraction method and locating LP A by combining texture features and color
features respectively.[12]
Shih-Jui Yang et.al.(2013) presented a robust and real-time LPD system to differentiate the strength and density of
vertical edges between the license plate region and non-plate regions. The proposed LPD system consisting of 2-level
2D Haar Wavelet transform and Wiener-deconvolution vertical edge enhancement not only can highlight the
irregularly-distributed vertical edges of the license plate region, but also can suppress the regularly-interlaced ones of
the non-plate regions, like radiator grilles or trench grates. [13]
Singh Gurjinder Pal(2012) Described Vehicles are identified by reading their number plate and then retrieving the
information from the record based on the number plate contents. The system becomes complicated when there is large
number of vehicles being traced at different locations. An automatic visual vehicle number plate identification and
management system is required that can grab the image of the moving/stationary vehicle’s number plate, extract the
contents from the same and then retrieve the details of the vehicle under surveillance. [14]
Md. Mahbubul Alam Joarder et.al(2012) Bangla automatic number plate recognition (ANPR) system using
artificial neural network for number plate inscribing in Bangla is presented in this paper. This system splits
into three major parts- number plate detection, plate character segmentation and Bangla character recognition. The
Bangla character recognition is implemented using multilayer feed-forward network. According to the experimental
result,
the performance of the proposed system on different vehicle images is better in case of severe image
conditions.[15]
Chirag Patel et.al(2013) Automatic Number Plate Recognition (ANPR) system are based on different methodologies
but still it is really challenging task as some of the factors like high speed of vehicle, non-uniform vehicle number plate,
language of vehicle number and different lighting conditions can affect a lot in the overall recognition rate.[16]
P.Vijayalakshmi et.al(2012) described new vehicle identification technique consists of vehicle detection, plate
localization and character recognition. Here, Genetic algorithm (GA) is employed at two levels: for detecting vehicle
from traffic image and recognizing character from the number plate. Detection is based on contour and shape
information. GA controls the window size to capture each vehicle in a separate widow.[17]
YiQing Liu et.al(2011) described a license plate recognition system based on neural networks was designed and
developed. The system used a neural network chip to recognize license plates. The chip combined video image
processing module with neural network module by using equalized image processing algorithm and network
classification algorithm. A set of interface circuit was developed for implementing license-plate-number
recognition.[18]
Muhammad Tahir Qadri et.al (2009)- defined Automatic Number Plate Recognition (ANPR) is an image processing
technology which uses number plate to identify the vehicle. The objective is to design an efficient automatic authorized
vehicle identification system by using the vehicle number plate. Normally, the data base is maintained in different
Volume 4, Issue 5, May 2015
Page 375
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 4, Issue 5, May 2015
ISSN 2319 - 4847
fields using the MS Access. The data base access must be made to the authorised user and must be password protected.
The modification rights must be reserved with the administrator and that too in different tiers. [19]
3.ALGORITHM
In order to implement the proposed algorithm; design and implementation has been done in MATLAB using image
processing toolbox. Table 4.1 is showing the various images which are used in this research work for experimental
purpose. Images are given along with their format and size. All the images are of different kind and also identify the
number plate for each image. Following steps are involved in designing of the proposed algorithm:
Step 1: Select Input of the original (RGB) image from computer memory into current program. Any given digital
image is represented as an array size M*N pixels.
Step 2: RGB to Gray level conversion.
Step 3: Image Enhancement and Binarization .
Step 4: Segmentation of texts and Hopfield Neural Network Training.
Step 5: Pixel Pattern for Hopfield NN for character identification
Step 6: Identification of Texts and store in a file.
Step 7: Complete no. Plate identification and comparing with the data base to get the owner’s information.
4.TRAINING WITH HOPFIELD NEURAL NETWORK
When testing simple distinct patterns the network performed well, correctly identifying each pattern. However, as
expected, as the patterns increased in similarity, the network often returned incorrect results. These additional states
(local minima) dramatically affected the network's ability to associate an input with the correct pattern. Following
equation is used for input character pattern identification:
P = 0.15n (where P = number of patterns and n = number of neurons). The network was subjectively tested using
numeric digits. These tests involved training the network with binary patterns that resembled a numeric digit followed
by a testing phase where numeric digits to be tested, were hand drawn using the computers mouse. The Hopfield
network correctly identified each number and returned the correct character
5.RESULTS
The presented algorithm has been tested on no. of images of vehicle registration plates. The snap shots shown in below
figures show the results. The characters are segmented and recognized at good accuracy. The rotation invariency has
also been tested for the same vehicle plates. A data base of 100 vehicles is generated for testing of the algorithm. The
data from the data base can be fetched when the number is extracted from the number plate. The data base is designed
using the following fields:
 Name of the owner
 Vehicle Number
 Address
 Model No.
 Chassis No.
 Engine No.
Table 1:Recognition Time (seconds)
Test Image
Correlation Based
DWT/NN Based
Proposed
Test-001
1.34
1.22
0.090
Test-002
1.34
1.20
0.091
Test-003
2.34
1.32
0.095
Volume 4, Issue 5, May 2015
Page 376
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 4, Issue 5, May 2015
Image1:Test_001
Image2:Test_002
Image4 : Input Image Test_001 and Resultant Output
ISSN 2319 - 4847
Image 3: Test_003
Graph1: Graphical Comparison – Recognition Time
Figure 2:Analysis and information of Number plate 1
Volume 4, Issue 5, May 2015
Page 377
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 4, Issue 5, May 2015
ISSN 2319 - 4847
Graph2 : Graphical Comparison - Accuracy
Image 5: Input Image Test_002 and Resultant Output
Image6: Input Image Test_003 and Resultant Output
Volume 4, Issue 5, May 2015
Page 378
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 4, Issue 5, May 2015
ISSN 2319 - 4847
Figure 2:Analysis and information of Number plate 2
Figure 3 :Analysis and information of Number plate 3
6.Conclusion
The proposed technique is applied and tested on different images and compared with respect to recognition time and
accuracy. Extensive research has been made in this proposed work to expand an accurate number plate recognition
system. Hopfield Neural Network generates less time and remarkable accuracy for pattern recognition as compared to
the previous methods. This indicates the speed of recognition is improved a lot. Further, the work may be taken out for
any other special characters that may be present on the number plate. However, the prime challenge that remains is the
exact location of the number plate. The position and style of number plate is always prone to vary and the algorithm
may have to be customized depending upon the requirement.
Volume 4, Issue 5, May 2015
Page 379
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 4, Issue 5, May 2015
ISSN 2319 - 4847
References
[1]. Gurjinder Pal Singh and 2Navneet Bawa,” License Plate Recognition using Hopfield Neural Networks,
International Journal of Modern Trends in Engineering and Research (IJMTER)Volume 02, Issue 03, [March 2015]
[2]. Singh Gurjinder pal et al . (2015 License Plate Recognition using Color based Segmentation and Neural
Networks”, International Journal of Scientific & Engineering Research, Volume 6, Issue 1, January-2015 594
ISSN 2229-5518.
[3]. Nandan More and Bharat Tidke,” License Plate Identification Using Artificial Neural Network and Wavelet
Transformed Feature Selection”, 978-1-4799-6272-3/15/$31.00(c)2015 IEEE.
[4]. Jitendra Sharm , Amit Mishra, Khushboo Saxena, Shiv Kumar,” A Hybrid Technique for License Plate
Recognition Based on Feature Selection of Wavelet Transform and Artificial Neural Network”, 2014 International
Conference
on
Reliability,
Optimization
and
InformationTechnology-ICROIT,978-1-4799-29955/14/$31.00©2014 IEEE.
[5]. Ayman Rabee and Imad Barhumi ,”License Plate Detection and Recognition in Complex Scenes Using
Mathematical Morphology and Support Vector Machines”, IWSSIP 2014, 21st tInternational Conference on
Systems, Signals and Image Processing, 12-15 May 2014, Dubrovnik, Croatia.
[6]. Cosmo H.Munuo1, Dr. Michael Kisangiri2,” Vehicle Number Plates Detection and Recognition using
improved Algorithms: A Review with Tanzanian Case study”, International Journal Of Engineering And
Computer Science ISSN:2319-7242 Volume 3 Issue 5, May 2014, Page No. 5828-5832.
[7]. IEEE, Mahmoud Ibrahim, Mohamed Shehata, “Automatic License Plate Recognition (ALPR):A State-of-theArt.Review”,IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 23,
NO. 2, FEBRUARY 2013 311
[8]. Firas Mahmood Khaleel #1, Siti Norul Huda Sheikh Abdullah #2, Muhamad Khuzaifah Bin Ismail #3,” Proc. of
the IEEE International Conference on Smart Instrumentation, Measurement and Applications (ICSIMA)” ,26-27
November 2013, Kuala Lumpur, Malaysia.
[9]. GuanXuezhong1,Liushuang1,LiZhi2,LiYunjiang3,SongTaolue1,LiuYang1 ,”Research on Automatic License Plate
Recognition Technology “2013 Third International Conference on Instrumentation, Measurement, Computer,
Communication and Control, 978-0-7695-5122-7/13 $26.00 © 2013 IEEEDOI 10.1109/IMCCC.2013.16
[10]. Bing-Fei Wu, Hao-Yu Huang ,Tsu-Tian Lee ,Chao-Jung Chen ,“Degraded License Plate Recognition System for
Town Buses on Highway”, 978-1-4673-5200-0/13/$31.00 ©2013 IEEE.
[11]. Jian Yang1, Bin Hu2 ,JieHan Yu3, Jianqi An 4, Gang Xiong,” A License Plate Recognition System Based on
Machine Vision”, 978-1-4799-0530-0/13/$31.00 ©2013 IEEE.
[12]. Yingjun Wu, Shouxun Liu and Xuan Wang,” License Plate Location Method Based on Texture and Color”, 9781-4673-5000-6/13/$31.00 ©2013 IEEE.
[13]. Shih-Jui Yang, Jia-Bin Jiang, Ming-Kan Wu, Chian C. Ho#,” Real-Time License Plate Detection System with 2level 2D Haar Wavelet Transform and Wiener-Deconvolution Vertical Edge Enhancement”, 978-1-4799-04341/13/$31.00 ©2013 IEEE
[14]. Chirag Patel , Dipti Shah, PhD. G. H. Patel ,Atul Patel, PhD.,” Automatic Number Plate Recognition System
(ANPR): A Survey, International Journal of Computer Applications (0975 – 8887) Volume 69– No.9, May 2013
[15]. Gurjinder Pal Singh, Chhailadeep Kaur,” Automatic Vehicle’s Number Plate Identification using Image
Segmentation and Comparison Techniques”, Volume 1 Issue 3, December 2012
[16]. Md. Mahbubul Alam Joarder, Khaled Mahmud, Tasnuva Ahmed, Mohsina Kawser, and Bulbul Ahamed ,”Bangla
Automatic Number Plate Recognition System using Artificial Neural Network” ATST-30224014©AsianTransactions.2012.
[17]. P.Vijayalakshmi.,M.Sumathi,” Design of Algorithm for Vehicle Identification by Number Plate Recognition”,
978-1-4673-5584-1/12/$31.00©2012 IEEE
[18]. Muhammad Tahir Qadri,Muhammad Asif,” AUTOMATIC NUMBER PLATE RECOGNITION SYSTEM FOR
VEHICLEIDENTIFICATION USING OPTICAL CHARACTER RECOGNITION”, 978-0-7695-3609-5/09 $25.00
© 2009 IEEE ,DOI 10.1109/ICETC.2009.54.
[19]. YiQing Liu and Dong Wei Ning Zhang and MinZhe Zhao,” Vehicle-License-Plate Recognition Based on Neural
Networks” 978-1-61284-4577-0270-9/11/$26.00 ©2011 IEEE
AUTHOR
The 1author is pursuing his M.Tech. (CSE) thesis work in image processing from ACET, Amritsar
(Punjab) India. His field of interest is in image processing based application development.
Volume 4, Issue 5, May 2015
Page 380