39

ISSN(Online): 2584-1025
MIT UNIVERSITY’S – ABHIVRUDDHI JOURNAL

VOL.1 (01) , JUNE , 2021 pp 39-41

Software Simulation of ANPR System using

MATLAB

Shrinivas Ajit Deshpande Abhishek Kumar Kashyap Kevin Rajesh Patel

Mechanical Department Mechanical Department Mechanical Department
MIT ADT University MIT ADT University MIT ADT University
Pune, India Pune, India Pune, India
deshpande.shrinivas09@[Link] [Link]@[Link] kevinpatel830@[Link]
Jai Amit Gandhi Nitanshu S Baviskar l
Mechanical Department Mechanical Department
MIT ADT University MIT ADT University
Pune, India Pune, India
[Link]@[Link]

MATLAB has commands integrated with it which help

Abstract—This paper highlights the importance and
in Image Processing [6]. If the number plate is matched
enhancement in the technology of the Automatic number with the Pre-occupied database, access is granted.
plate recognition system (ANPR) system and its future Otherwise, it’s denied [7]. The increasing number of
applications for universities, large apartments, andhighly accidents day by day makes it important for ANPR systems
restricted areas. The ANPR system uses the concept of (photo and video) to be used widely [8]. The amount of
image filtering and character recognition technology to light present in the environment where is ANPR system is
extract the characters from vehicle’s number plate and present needs to be monitored so as to get a better image
correlate them with the pre-occupied database. This for recognition [9]. As RGB images have a lot of data
system will work as an access grant system when the captured in them, it is important to convert them into
vehicle’s number plate matches with the pre-occupied grayscale for faster recognition [10].
database of the system, and then access will be granted.
Till now, we can see that the output for the Automatic
Otherwise, the access will be denied and the system will Number Plate Recognition System is saved in a text file or
send an alert message to the security guard or the directly in the command window of MATLAB. Here in this
concerned admin via E-mail or SMS. A low-cost model of paper, we have showcased the system to be able to send an
this system can be very well applied to apartments and SMS or an E-mail to the user. This research paper focuses on
societies where the problem of vehicle security in the the importance of using ANPR technology for universities
parking lot, maintaining the record of residents’ vehicles, and larger apartments or residential areas where security and
and keeping a watch on suspicious intruders in societies maintaining records of every vehicle manually is a tedious
can be solved. task, especially in a country like India. This simulation is
developed in MATLAB software which is easy to use and has
Keywords—ANPR, Image recognition system, MATLAB, good compatibility with cameras. Ease of use
Number Plate recognition
II. METHODOLOGY
I. INTRODUCTION
In the early 90s, the Automatic number plate recognition A. Image Acquisition
system (ANPR) was used as an integral part of the field of The first step in the process is to acquire the image. In the
security and law enforcement in the UK. In the early 90s, this suggested method, we will take an input image from the user
technology only used primary cameras to capture the image to aid the recognition process of the number plate, as shown
of suspected vehicle number plates for carrying out a clear in Fig. 1.
investigation. However, the development in this technology
increased when the primary cameras were replaced by the
digital cameras having a higher range and wide-angle [1]. A
synergy of data science and machine learning algorithms can
be used to enhance the ANPR system to function in the real
world. As the population increased over the period of time in
India and abroad, this technology has become the need of an
hour to manage traffic and also for the security and law
enforcement. In essence, it consists of a powerful digital
camera having a wide range that has the ability to capture Fig. 1. Image Acquired from Input
images of the vehicle number plate in real-time [2]. When the
image of the vehicle’s number plate is captured, the system
will segregate characters in the number plate by using a
character recognition algorithm [3], when the characters inthe
number plate will be checked with a pre-occupied large
database with correlation [4]. OCR, i.e., Optical Character
Recognition, plays a crucial role in this type of system [5].
 40
ISSN(Online): 2584-1025

B. Pre-Processing Of Number Plate C. Image Filtration

The acquired input images are in the RGB format. For The target is successfully found, but it is the target with
faster processing speed, they need to undergo a conversion some extra characters that has been found. Any entities in
from RGB to Greyscale. The RGB image is hence converted connection to the image border must be removed. This is how
into greyscale using the rgb2gray() function. the actual region of the plate is detected.

The RGB colour group consists of three primary colours, D. Character Segmentation
that is Red, Green and Blue. Greyscale, on the other hand, is
Segmentation is a crucial process in the NPR process, as
composed of two main colours that are black and white.
subsequent processes rely on it. If this stage fails, a character
Therefore, this aids in quicker processing. This can be seen in
may not be recognized properly that is, the character may be
Fig. 2.
divided into two or even merged to form a single character.
To avoid this, we use the bounding box technique. The
bounding box is used to measure the properties of the target
region. Once a bounding box is created for each and every
character in the image of the number plate, it becomes a
separate entity for recognition of the number plate as shown
in Fig. 3.

Fig. 2. RGB to B/W converted Image

Identifying the size of the plate is the first step in the
process of recognising a vehicle number plate. Since number
plates often have a rectangular shape, it is important to
recognise the edges of the plate. The mathematical
morphology method is utilised to find the region, and the
Sobel operator is used to get the threshold value that finds the
regions that stand out due to their strong edges and large edge Fig. 3. Bounding Box Implementation
variance.
The image forms a binary gradient mask as a result of E. Character Recognition Using Template
which distinct, sharp contrast lines are discernible. The target The process of converting text images into characters uses
image's outline is not easily distinguished by these lines. character recognition. Now, utilising template matching
There are still a lot of gaps and noise around the lines that arrays, the number plate identification process compares each
surround the item in the gradient mask as compared to the character to the entire alphanumeric database. The template
original Image. picture moves to every available point on the bigger source
image throughout the matching process, and as a result, a
If the Sobel image is improved with linear structuring
numerical index is computed to show how well the template
components, this linear gap is eliminated. Matrix
matches the image in that position. Pixel by pixel, the
representations of the structuring components are used to
matching procedure takes place. Fig. 4 displays the image
measure the form of a picture. Matrix representations are a
templates.
characteristic of specific structures and characteristics. Other
image processing processes are performed using this. The
vertical structuring element is used first to improve the binary
gradient mask, and then the horizontal structuring element.
The MATLAB toolbox provides a pre-defined function
imfill() which fills any overlooked holes in the binarized
image. The enhanced gradient mask distinguishes the outline
of the cell quite appropriately but does leave unwanted holes
(impurities) inside the boundary. The image is scanned for
any impurities, and if found, they are filled in accordance
with the pixels in the background.
 41
ISSN(Online): 2584-1025

The system thus successfully recognizes the all the

characters from the Input image by neglecting the
surrounding environment. The system can segment each and
every character individually for recognition. It displays the
number plate of the desired vehicle and also sends a mail to
the registered mail id that is present in the program so that
it cannot be changed by the user.

III. CONCLUSION
In this paper, automatic vehicle identification and
reporting system from vehicle license plates are presented.
The system uses numerous image processing and character
recognizing techniques for analyzing number plates. That is
implemented using the MATLAB platform and gives results
after processing images. The image processing uses the tools
and commands provided in MATLAB software. Thisprogram
developed in MATLAB first extracts the number plate from
the image provided and then recognizes the character using
MATLAB commands and it’s saved in the array, which forms
the output of the program, which can be sent to the user via
text file or E-mail or SMS.
The real-world performances are to be tested, as of now,
Fig. 4. Template used for Template Matching only virtual simulations have been performed, but the results
are great. The integration with e-mail gives our system the
ability to instantly notify the security team about unknown
F. MATLAB Results
vehicles entering an area.
The ultimate outcome of ANPR system is shown in Fig. 5
and Fig. 6 below:
REFERENCES
[1] Rahini Bhat,Bijendra Mehandia, “Recognition of Vehicle Number
Plate”, IRJST, vol2, issue8, Aug2014.
[2] Mandeep Kaur Ahuja,[Link] Singh, “Hand Gestures
Recognition using PCA”, IJCSET, vol5, issue7.
[3] Mithai Ovidiu, “Performance Study based on MALAB Modelling for
Hybrid Vehicles”, IJCA, vol99, issue12, Aug 2014.
[4] Muhammad Tahir Quadri,Muhammad Asif, “ANPR Systems for
vehicle identification using optical character recognition, “IRJET,
vol5, issue5, May 2018.
[5] Narendra Singh Tomar,Prakhar Sachen,Pranav Mittal, “Vehicle
Number Plate Detection using MATLAB, “IRJET, vol5, issue5, May
2018.
[6] Shilpi Chauhan, Vishal Shrivastav, “MATLAB based vehicle number
Fig. 5. Resultant Image plate recognition”, “IJCIR, vol3, issue 9,2017.
[7] Bhawana Tiwari, Archana Sharma, “Automatic Vehicle Number Plate
Recognition Using MATLAB, “IRJET, vol3, issue5, May2016.
[8] Amit Kumar Parida,SH Mayuri. “Recognition of vehicle using
MATLAB”, “IJFCST, vol4, number3, 2014.
[9] Rathore, M., & Kumari, S. (2014). Tracking number plate from vehicle
using Matlab. International Journal in Foundations of Computer
Science & Technology (IJFCST), 4(3), 43-53.
[10] FAOUR, REHAM & Shanwar, Bassel & Zarka, Nizar. (2016).
RECOGNITION OF VEHICLE NUMBERPLATE USING
MATLAB. 10.13140/RG.2.1.1459.8640.
Fig. 6. Image of E-mail