2017 IEEE International Conference on Smart Technologies and Management

for Computing, Communication, Controls, Energy and Materials (ICSTM),

Veltech Dr.RR & Dr.SR University, Chennai, T.N., India. 2 - 4 August 2017. pp.227-231.

Entry and Exit Monitoring using License Plate Recognition

A. Yovan Felix and A. Jesudoss and J. Albert Mayan

Assistant professor,
Department of Computer Science and Engineering, Sathyabama University, Chennai.
Email: [email protected], [email protected], [email protected]

Abstract— In the transport and traffic management system, easily by several circumstances and if it is automated the
tracking of vehicles on road is a prime importance vehicles are data are saved and it can be easily printed or a soft copy can
identified by reading their number plate, the existing system be forwarded to parents via mail.
becomes complicated when there is a large number of vehicles
being traced at different locations and therefore automatic This system started to work with two processes namely
license plate recognition ALPR is used to extract the contents
of the number plate. The image can be fetched by using CCTV
Automatic process and Manual Process, the manual process
cameras with an MATLAB image processing tool. Besides are eliminated by an employee who needs to maintain it and
these can be used for a surveillance purpose to facilitate the is often difficult to comply with regulation but an automated
operations at the toll of many gated communities the system is valuable for ensuring compliance and regulation
identification of vehicles are done through vehicles through proof of attendance.
motion locating the license plate and accurately identifying the
numbers in the license plate. Here we are going to use
Adaptive Histogram Equalization AHE, Active contour
method for region separation Optical Character Recognition
OCR to recognize the characters and Deep Neural Network.
DNN is used for classification and extracting the text as an
alphanumeric characters and comparing the text with the
predefined table created in MYSQL server and changing the
status accordingly.

Index Terms— ALPR CCTV, License plate, alphanumeric

characters, AHE, OCR, DNN, MATLAB, MYSQL.

I. INTRODUCTION

The Organizations of all standards maintains a general

system for the vehicles for their needs some organizations
keeps a detailed report of the attendance issues. An
attendance system provides many benefits for the
organization. There was a time the attendance of the
vehicles and the employees use to mark at registers
manually for each vehicles entering inside the organization.
Now in this automated world of technology, the technology
needs to play a vital role in each and every field. The Fig. 1. Block Diagram of license plate recognition
Vehicle attendance monitoring made lives simple and
elegant the vehicle attendance monitoring system makes the
employee way of marking attendance of a vehicle a piece of
cake.

In universities and colleges, parents are never being

informed of the vehicle entering inside the campus or not. If Fig. 2. Image of license plate in English format with high
it is done manually the record or registers can be exploited luminance - Image of the standard number plate

978-1-5090-5905-8/17/$31.00 ©2017 IEEE 227

2017 IEEE International Conference on Smart Technologies and Management for Computing, Communication, Controls, Energy and Materials

II. LITERATURE SURVEY IV. PROPOSED ALGORITHM

Nauman Salim, Hassam Muazzam, H.M.Tahir proposed The algorithm of the proposed system is given below:
“Automatic License plate recognition Using Extracted
a) Input video recording CCTV camera
features “ advantage of this concept is execution time was
extremely fast it is stated as 0.5 seconds and the b) Trigger camera as the number plate caught in the frame
disadvantage is accuracy level is very low. c) Get the image of the number plate
Pooja Sharma, Shanu Sharma proposed “Image Processing d) Convert the image to a grayscale image
Based degraded Camera Captured Document Enhancement e) Apply AHE to increase the saturation level and median
For Improved OCR Accuracy” Advantages of this concept filter to reduce the noise of an image
is Noisy and Blurred images are extracted and the
Disadvantage is extraction is poor with low resolution. f) Convert the image to a black and white image
g) Then convert the image to a binary image
Soumitro Swapan auddy, Kanishka Tyagi, Son Nugen
proposed”Discriminant Vector Transformations in Neural h) Remove the edges (horizontal and vertical) apart from
Network Classifiers” advantage of this concept is the boundary conditions using
classification perfection is accurate and the Disadvantage is i) Using OCR identify the string presented in the license
execution time is slower. plate
Younghui jia, Thomas Gannot, Jaffar Sanie, j) Using DNN classify and print then store it as a string
Proposed”Design Flow Of Vehicle Plate Reader based on k) Compare the string with the predefined data in the table
RGB color Extractor” advantage of this concept is OCR
accuracy is well optimized to 98.5% and the disadvantage l) Then print the Status
of this concept is execution time is slow.
V. MODULES
The existing system suffers from the following drawbacks:
The proposed work consists of the following modules:
a) Only CCTV based monitoring.
a) Camera module
b) Only OCR based License plate recognition.
b) Preprocessing with AHE and median filter
c) Manual method of attendance marking and maintaining
c) Scaled region detection with Active Contour method
in records
d) Segmentation using OCR

III. PROPOSED SYSTEM e) Edge detection

f) Classification using Deep Neural Network
The proposed system offers the following advantages:
g) Matching values with the predefined Table
a) Digital cameras are cheap, versatile, and handy
h) Printing the status
b) The paper deals with OCR and Active Contour method
for number plate recognition A. Camera Module
c) Active contour is widely used for color detection and An CCTV camera with the video recording capability is
region separation given as a input in order to record the moving vehicles, the
d) Deep Neural Network DNN method is used for camera is fixed at particular place so that it can record or
classification with the trained table capture the License plate in a frame whereas the camera
triggering is manually operated by an employee.
e) The median filter is used for noise reduction
f) The result after classification is alpha numeric string
g) The text is stored in a database and the stored text is
matched with the predefined table created
h) The matching of the text is done with JAVA with
MYSQL queries Fig. 3. Image captured from one of the frames from the live
running camera

228
 Entry and Exit Monitoring using License Plate Recognition

B. Preprocessing with AHE and Median Filter D. Segmentation by using OCR

Preprocessing evolves with the conversion of RGB image to After the process of detecting the scaled image, the subject
a gray scaled image in order to reduce luminance. that is presented in License plate has to be converted to a
binary image so that the characters presented in the license
plate can be easily identified by Optical Character
Recognition OCR.

Fig. 4. Conversion of RGB image to a Gray scaled image

Adaptive Histogram Equalization (AHE)— After the

conversion of RGB image to a gray scaled image, apply
Adaptive Histogram Equalization in order to increase Fig. 7. The above given image is the binary image and the
the saturation of color. characters in the License plate are identified by OCR
Median filter—During the equalization there is a E. Edge Detection
possibility for minor noise that is said to be created in
the image in order to filter noise the median filter is In order to remove the edges that is horizontal and vertical
used. edges which is apart from the boundary lines are removed
because it may give some complexity to the algorithm

Fig. 5. The saturated and denoised image

Fig. 8. Edged image
C. Scaled Region Detection using Active Contour
Method F. Classification using Deep Neural Network

After getting the denoised image we need to convert the

image into a Black and White image, after getting the black
and white image the Active Contour Method is used to
separate the scale region which means the foreground is
noted by separating the background

(a) Black and White Image converted from the denoised

image

(b) Scale region detected by Active Contour method Fig. 9. The above image is the DNN performing the
classification of characters presented in the license plate
Fig. 6. Active contour method Based scaled region detection

229
2017 IEEE International Conference on Smart Technologies and Management for Computing, Communication, Controls, Energy and Materials

The classifier algorithm DNN is used to process most

accurate result after the OCR results, it is used to match the
characters with the predefined trained dataset, it handles all
sorts of alpha numeric fonts even hand written English fonts
are also readable by using the Deep Neural Network
classifier.

Fig. 13. Showing the Status of overall vehicle entry and exit
details

As soon as the alpha numeric string of the license plate

Fig. 10. The output of the DNN printing the alpha numeric number gets matched with the predefined table then it
characters changes the status of the particular row. Likewise whatever
the number the MATLAB program generates the changes
are done in the status only if the same alpha numeric string
are already defined in the table.

VII. RESULT AND FUTURE WORKS

It can be said as the extraction of characters from the license

plate and changing the status accordingly with comparing to
the table works with an 100% accuracy level. This system
can be widely used among the gated communities and other
Fig. 11. The output of the DNN is opened in a notepad in (.txt) organizations in order to monitor their campus vehicle and
format and saved to show their exact status of their campus
G. Matching the Values with the Predefined Table
Initially turn on the main server program that should be REFERENCES
always running to match the values that is generated by
[1] Nauman Salim, Hassam Muazzam, H.M Tahir
DNN classifier. Create the table of rows and columns as per
“Automatic License Plate Recognition Using
our wish by using Queries that are used in MYSQL save all
Extracted Features” Computational and Business
the number plate alpha numeric characters inside the table,
Intelligence (ISCBI), 2016 4th international
whenever a number plate is recognized it gets stored as a
symposium on 5-7 September 2016 INSPEC
text the java combined with MYSQL code fetches it and
Accession Number: 16466966 DOI:
matches with the table
10.1109ISCBI.2016.7743288.
H. Printing the Status [2] Pooja Sharma, Shannu Sharma “Image Processing
Based degraded Camera Captured Document for
Improved OCR accuracy” Cloud System and Big Data
Engineering (Confluence), 2016 6th International
Conference on 2016 INSPEC Accession Number:
16140689
[3] Albert Mayan.J, Surya.B, Pranoy Prabhakar,
Princekumar,” Department–Student Library Using
Twig Pattern Query Processing Over Admin-User
Login Privilege”, Pak. J. Biotechnol. Vol. 13, Pp. 489
- 493,(2016)
[4] Mayan J.A., Ravi. T, “Test Case Optimization Using
Hybrid Search Technique”, ACM International
Fig. 12. Displaying the status for particular vehicle number Conference Proceeding Series, ACM New York, NY,

230
 Entry and Exit Monitoring using License Plate Recognition

USA ©2014, ISBN: 978-1-4503-2908-8, doi:10.1145/ [10] B.Bharathi and Mahesh kumar (2016), ‘Non invasive
2660859.2660906 BG scrutinizer system’, Global Journal of Pure and
[5] Soumitro Swapan auddy, Kanishka Tyagi, Son Nugen Applied Mathematics, vol.12, No: 8, pp. 5123 – 5125
“Discriminant Vector Transformation in Neural [11] Surya.V, J. Albert Mayan,” A Secure Data Sharing
Network Classifiers” Neural Networks (IJCNN), 2016 Mechanism In Dynamic Cloud By Using KP-ABE”,
International joint Conference on 24-29 July 2016 Research J. Pharm. and Tech. 2017; 10(1): 83-86
INSPEC Accession Number: 16446866 DOI:10.1109 [12] K. Mohana Prasad, Dr. R. Sabitha “Meta Physical
/IJCNN.2016.7727415. Algorithmic Representation for Flawless Clustering”
[6] Sudhakar.M, Mayan J.A., Srinivasan.N, “Intelligent Journal of Theoretical and Applied Information
data prediction system using data mining and neural Technology (JATIT), Volume 76, NO 1,PP 82-87.
networks”, Advances in Intelligent Systems and [13] S.Karthikeyan, S.Jayashri, “Proficient Data Gathering
Computing, March 2015 by Mobile Agent using Jade”,i-manager’s Journal on
[7] Yonghui Jia, Thomas Gannot, Jafar Sanie “Design wireless communication networks,Vol.1, No. 2, 2012,
Flow of Vehicle License Plate Reader Based on RGB
pp. 32-40.
Color Extractor” Electro Information Technology
[14] Albert Mayan J, Kumar Akash Deep, Mukesh Kumar,
(EIT), 2016 IEEE International Conference on 19-21
Livingston Alvin, Siva Prasad Reddy, “Number plate
May 2016 INSPEC Accession Number: 16212177
recognition using template comparison for various
DOI: 10.1109/EIT.2016.7535290.
[8] B Vaishali, J Jabez, “Traffic Surveillance Of Objects fonts in MATLAB “, IEEE International Conference
Utilizing The Lower Bandwidth”,Research Journal Of on Computational Intelligence and Computing
Pharmaceutical, Biological And Chemical Research, ICCIC 2016, PP: 1-6, DOI:
Sciences,Volume 6, Issue 3, pp:999-1004,2015 10.1109/ICCIC.2016.7919542
[9] Mary Posonia. A, V.L.Jyothi,”Efficient XML [15] Jesudoss A, Subramaniam N.P,”Securing cloud-based
Keyword search using H-Reduction factor and healthcare information systems using enhanced
Interactive Algorithm”,International Review on password-based authentication scheme”, Asian Journal
Computers and Software(IRECOS),Vol-9,N-12,pp- of Information Technology,Vol 15, Issue 14, pp: 2457-
2022-2030,2014 2463,2016.

231