Automatic Number Plate Recognition
System
SIGNALS AND SYSTEM- ECE1004
AMAN RAJ GUPTA
(18BEC0321)
ECE100
Abstract:
4
The main objective of our project is to
recognize and display the Registration
numbers of vehicles which are scanned
by the camera.
As number plate is one of the unique
identifications used by the government
officials to recognize a vehicle.
Automatic number plate recognition is
one of the best ways to fulfil this
demand. The number written on the
plate is can be used by the police
officials to stop the criminal activities
taking place. Traffic police can use
Automatic number plate recognition to
keep check on vehicles which violate
the traffic rules and are driving too
fast.
This system can also be used for
surveillance, traffic management. This
system can be put to surveillance 24/7
Ch.Saikiran Shabd Prakash
(18BEC0278) (18BIS0031)
even without the availability of human
force.
In this project is basic application of
image processing by virtue of which
we can capture image convert it into
the form readable by computer
followed by noise reduction and
various filtering techniques which
include sobel edge detector. Sobel
edge detector acts as 3X3 mask
creating an image emphasizing edges.
Further the advantages and
disadvantages of the process involved
is discussed.
Keywords:
Automatic Number Plate Recognition
(ANPR), MATLAB, Conversion,
Segmentation, Character Recognition.
1
Introduction: Block Diagram
The Automatic Number Plate
Recognition System is an age-old
technology which is now getting
enhanced day by day by various
algorithms developed by signal
scientists and now artificial
intelligence has come into the sector
with the integration of Internet of
Things. Now days this system is
mostly used for identification of
various types of vehicles. These
systems were previously used by
government officials to identify
various criminals.
This system uses basic principle like
Image Reading, Image Conversion,
Image Resizing, Image Filters like
(erode, fill, dilation), Image Morphing
Procedure:
and Image Edge Detection (Sobel Image Reading:
Feldman Operator), Image
Comparison, Template Reading, We input the raw image which is
Image Processing. actually an RGB image (analog input)
which is considered for conversion.
This system is 60%-70% accurate.

2
Gray Code Conversion:
This RGB image is converted to Gray
Scale code which is again binarized to
8-bit Gray code.
Filtering:
After conversion the image is filtered
using paper and salt noise reduction
which consists of eroding, filling and
dilating. The corresponding 3*3 masks
get eight neighbouring values.
3
License Plate Detection:
This step is the most important step in
the detection of this system and
algorithm which is based on the belief
that license plates are rectangular in
shape so mathematical morphology is
used for detecting the region of the car
where the number plate is fixed and no
other area gets highlighted due to this.
Now by using the Sobel Feldman
Operator we are using which
highlights the edge area using high
edge variance formula.
Segmentation:
This step is again the next most
important step of where the license
plate is segmented using the character
detection technique which is mainly
given by their vertical projection.
 Character Recognition: Display Output:
This step we compare different Then each output is displayed and each
characters which are usually step is shown which renders and reads
combinations of Numbers and and then finally displays the
Alphabets of varying length. These characters.
characters are compared using the
database of templates given by the use
of unique indexes. Which are
numbered so as to compare the no. of
end points.

Code:-
Plate_detection.m
close all;
clear all;

im = imread('Number Plate Images/image3.png');

im = imresize(im, [480 NaN]);
subplot(3,3,1),imshow(im),title('Resized Image')
imgray = rgb2gray(im);
subplot(3,3,2),imshow(imgray),title('Gray scale Image')
imbin = imbinarize(imgray);
subplot(3,3,3),imshow(imbin),title('Binary Image')
im = edge(imgray, 'sobel');
subplot(3,3,4),imshow(im),title('Edge Detection')
im = imdilate(im, strel('diamond', 2));
subplot(3,3,5),imshow(im),title('Dilated Image')
im = imfill(im, 'holes');
subplot(3,3,6),imshow(im),title('After filling holes')
im = imerode(im, strel('diamond', 10));
subplot(3,3,7),imshow(im),title('Eroded Image')

4
 jufgdhgigdfj
Iprops=regionprops(im,'BoundingBox','Area', 'Image');
area = Iprops.Area;
count = numel(Iprops);
maxa= area;
boundingBox = Iprops.BoundingBox;

for i=1:count
if maxa<Iprops(i).Area
maxa=Iprops(i).Area;
boundingBox=Iprops(i).BoundingBox;
end
end
im = imcrop(imbin, boundingBox);
im = imresize(im, [240 NaN]);
im = imopen(im, strel('rectangle', [4 4]));
im = bwareaopen(~im, 500);
[h, w] = size(im);
subplot(3,3,8),imshow(im),title('Resize the Bounded Box')

imshow(im);

Iprops=regionprops(im,'BoundingBox','Area', 'Image');
count = numel(Iprops);
noPlate=[];

for i=1:count
ow = length(Iprops(i).Image(1,:));
oh = length(Iprops(i).Image(:,1));
if ow<(h/2) && oh>(h/3)
letter=Letter_detection(Iprops(i).Image);
noPlate=[noPlate letter]
end
end

5
Letter_detection.m
function letter=readLetter(snap)
load NewTemplates
snap=imresize(snap,[42 24]);
rec=[ ];

for n=1:length(NewTemplates)
cor=corr2(NewTemplates{1,n},snap);
rec=[rec cor];
end

ind=find(rec==max(rec));
display(ind);

.
if ind==1 || ind==2
letter='A';
elseif ind==3 || ind==4
letter='B';
elseif ind==5
letter='C';
elseif ind==6 || ind==7
letter='D';
elseif ind==8
letter='E';
elseif ind==9
letter='F';
elseif ind==10
letter='G';
elseif ind==11
letter='H';
elseif ind==12
letter='I';
elseif ind==13
letter='J';
elseif ind==14
letter='K';
elseif ind==15
letter='L';
elseif ind==16
letter='M';
elseif ind==17
letter='N';

6
elseif ind==18 || ind==19
letter='O';
elseif ind==20 || ind==21
letter='P';
elseif ind==22 || ind==23
letter='Q';
elseif ind==24 || ind==25
letter='R';
elseif ind==26
letter='S';
elseif ind==27
letter='T';
elseif ind==28
letter='U';
elseif ind==29
letter='V';
elseif ind==30
letter='W';
elseif ind==31
letter='X';
elseif ind==32
letter='Y';
elseif ind==33
letter='Z';
elseif ind==34
letter='1';
elseif ind==35
letter='2';
elseif ind==36
letter='3';
elseif ind==37 || ind==38
letter='4';
elseif ind==39
letter='5';
elseif ind==40 || ind==41 || ind==42
letter='6';
elseif ind==43
letter='7';
elseif ind==44 || ind==45
letter='8';
elseif ind==46 || ind==47 || ind==48
letter='9';
else
letter='0';
end
end

7
Template_detection.m
A=imread('alpha/A.bmp');B=imread('alpha/B.bmp');C=imread('alpha/C.bmp');
D=imread('alpha/D.bmp');E=imread('alpha/E.bmp');F=imread('alpha/F.bmp');
G=imread('alpha/G.bmp');H=imread('alpha/H.bmp');I=imread('alpha/I.bmp');
J=imread('alpha/J.bmp');K=imread('alpha/K.bmp');L=imread('alpha/L.bmp');
M=imread('alpha/M.bmp');N=imread('alpha/N.bmp');O=imread('alpha/O.bmp');
P=imread('alpha/P.bmp');Q=imread('alpha/Q.bmp');R=imread('alpha/R.bmp');
S=imread('alpha/S.bmp');T=imread('alpha/T.bmp');U=imread('alpha/U.bmp');
V=imread('alpha/V.bmp');W=imread('alpha/W.bmp');X=imread('alpha/X.bmp');
Y=imread('alpha/Y.bmp');Z=imread('alpha/Z.bmp');

one=imread('alpha/1.bmp');two=imread('alpha/2.bmp');
three=imread('alpha/3.bmp');four=imread('alpha/4.bmp');
five=imread('alpha/5.bmp'); six=imread('alpha/6.bmp');
seven=imread('alpha/7.bmp');eight=imread('alpha/8.bmp');
nine=imread('alpha/9.bmp'); zero=imread('alpha/0.bmp');

letter=[A B C D E F G H I J K L M N O P Q R S T U V W X Y Z];
number=[one two three four five six seven eight nine zero];

NewTemplates=[letter number];
save ('NewTemplates','NewTemplates')
clear all

8
RESULT:

9
Application of Automatic
Number Plate Recognition
Automatic Number Plate Recognition
has a wide range of applications since
the license number is the primary, most
widely /accepted, human readable,
mandatory identifier of motor
vehicles.
It provides automated access of the
content of the number plate for
computer systems managing databases
and processing information of vehicle
movements.
Parking
One of the main applications of ANPR
is parking automation and parking
security: ticketless parking fee
management, parking access
automation, vehicle location guidance,
car theft prevention, "lost ticket" fraud,
fraud by changing tickets, simplified,
partially or fully automated payment
process, amongst many others.
10
Law Enforcement
Automatic number plate recognition is
an ideal technology to be used for law
enforcement purposes. It is able to
automatically identify stolen cars
based on the up-to date blacklist. Other
very common law enforcement
applications are red-light enforcement
and overspeed charging and bus lane
control.
Toll enforcement
It provide commuters with an efficient
and time saving opportunity, ANPR is
used at tolling stations in a wide range
of countries as a method of validating
vehicle types and payment. Rather
than waiting in queues or worrying
about the correct change, ANPR
systems have been set up to capture
number/license plates of those passing
through the toll plaza, allowing drivers
to go online or ring later and pay the
toll fee rather than having to stop with
cash.
CONCLUSION
The results are as expected but not to
up to the mark. The time of
compilation and processing is varying
depending on the size of image and the
resize function. Around 12 images
have been tested by us at different
angles but even if we increase the
range of resize function it is detecting
only of the characters. We were able to
find out that bigger images which have
other vehicles or writings in
background could not be processed
correctly. We weren’t able to do
extrapolation and interpolation. More
changes can be executed done on this
model. A GUI based software can be
developed using more complex code
and the existing algorithm.
11
REFERENCES
o Qadri, Muhammad Tahir, and
Muhammad Asif. "Automatic
number plate recognition system
for vehicle identification using
optical character recognition."
Education Technology and
Computer, 2009. ICETC'09.
International Conference on.
IEEE, 2009
o Kurdi, Moustafa M., et al.
"Lebanese automated number
plate reading based on neural
network recognition." Computer
Engineering Conference
(ICENCO), 2017 13th
International. IEEE, 2017.
o Panahi, Rahim, and Iman
Gholampour. "Accurate detection
and recognition of dirty vehicle
plate numbers for high-speed
applications." IEEE Transactions
on intelligent transportation
systems 18.4 (2017): 767-779.