VEHICLE NUMBER PLATE DETECTION

A report on Project based learning for DSP LAB

BACHELOR OF TECHNOLOGY

In
ELECTRONICS AND COMMUNICATION ENGINEERING
Submitted by

G. VAMSI KRISHNA - 21331A0447

G. NIKHILESWAR - 21331A0448

G. CHAITANYA - 21331A0449

Under the guidance of

Dr. G. Vimala Kumari

M.Tech., Ph.D
ASSOCIATE PROFESSOR

ECE

Department of Electronics and Communication Engineering Maharaj

Vijayaram Gajapathi Raj College of Engineering(A)
Approved by AICTE, New Delhi Permanently Affiliated to JNTU, Kakinada.
Re-Accredited by NBA and NAAC ‘A ’grade, listed u/s 2(f) and 12(B) of the UGC Act
1956 Vijayaram Nagar Campus, Vizianagaram, Andhra Pradesh-535005. 20232024

1|Page
 CONTENTS

TITLE PAGE
NO ABSTRACT Error! Bookmark not defined.

INTRODUCTION 4

METHODOLOGY 4

SOURCE CODE 5

REULTS 6

RESULT ANALYSIS 7

APPLICATIONS 7

DISADVANTAGES 8

CONCLUSION 9

REFERENCES 9

2|Page
ABSTRACT:

The exponential growth of vehicles on roads necessitates efficient and automated

methods for vehicle number plate detection to enhance traffic management, law
enforcement, and security systems. This abstract proposes an innovative approach
leveraging image processing techniques for the automatic detection of vehicle
number plates.

The proposed system consists of several stages, starting with image acquisition
from cameras installed at strategic locations. Preprocessing techniques such as
noise reduction, contrast enhancement, and image resizing are applied to improve
the quality of the acquired images. Following preprocessing, a series of image
processing algorithms are employed to isolate and extract the number plate region
from the acquired images.

The core of the system lies in the number plate detection algorithm, which utilizes
various techniques such as edge detection, morphological operations, and contour
analysis to accurately locate the number plate within the image. Once the number
plate region is identified, character segmentation techniques are employed to
isolate individual characters present on the plate.

Next, Optical Character Recognition (OCR) algorithms are employed to

recognize and extract alphanumeric characters from the segmented regions. This
stage involves the training of machine learning models on a dataset of characters
to enhance recognition accuracy. Post-processing techniques such as character
validation and correction are applied to refine the OCR results and ensure accurate
extraction of the vehicle number.

The proposed system offers several advantages, including real-time processing

capability, high accuracy in number plate detection and character recognition, and
scalability for deployment in various traffic and security applications.
Additionally, the system can adapt to different environmental conditions such as
varying illumination and weather conditions, thus ensuring robust performance in
real-world scenarios.

In conclusion, the proposed automated vehicle number plate detection system

presents a sophisticated solution to the growing challenges associated with traffic
management and security enforcement. By leveraging advanced image
processing and machine learning techniques, the system offers a reliable and
efficient means of automatically detecting and recognizing vehicle number plates,

3|Page
thereby contributing to improved road safety, traffic flow, and law enforcement
efforts.

INTRODUCTION:

The advent of image processing technology has revolutionized various domains,

including traffic management and security systems. One critical application of
image processing in this domain is the automated detection of vehicle number
plates. Traditional methods of manual inspection are time-consuming and
errorprone, necessitating the development of automated systems to streamline the
process. This introduction provides an overview of a proposed system that
leverages image processing techniques for the automatic detection and
recognition of vehicle number plates. By combining advanced algorithms with
machine learning models, the system aims to enhance efficiency, accuracy, and
reliability in traffic monitoring, law enforcement, and security operations.

METHODOLOGY:
The process includes processing the image. First the target image is get affected
with the binary filter which darkens the blacks and lightens the whites. Later it
get followed few more filters to remove the background objects into dark.

Now the number plate is clearly visible and it get cropped the potion need to be
know. The part is in black and white form.

A samples of 36 images are given as an input or loaded to the program ,which

contain the binary images of all 26 alphabets (A-Z) and numbers from 0-9.

The binary images are assigned with respected characters. Now the crop image
get compared with these 36 samples if any match get found the letter get printed
and the process continues till the last pixels. And the number is detected The
samples images :

4|Page
SOFTWARE USED :

1)MATLAB SOFTWARE

SOURCE CODE:
[path file] = uigetfile('/MATLAB Drive/test2.jpg','Select Image');
img = imread(fullfile(file,path)); figure; imshow(img);
title('Original Image'); img2 = rgb2gray(img);
imgX = edge(img2,'sobel','horizontal');
imgY = edge(img2,'sobel','vertical');
img3 = imgX|imgY; img3 = bwareaopen(img3, 700);
se = strel('line',2,90); img3 =
imdilate(img3,se); se = strel('line',2,0); img3
= imdilate(img3,se); img4 =
imfill(img3,'holes'); img4 =
bwmorph(img4,'thin',1); se = strel('line',5,90);
img5 = imerode(img4,se); img5 =
imerode(img5,se); img6 = immultiply(img2,img5);
verticalProfile = sum(img6, 2); row1 =
find(verticalProfile, 1, 'first'); row2 =
find(verticalProfile, 1, 'last');
horizontalProfile = sum(img6, 1); column1 =
find(horizontalProfile, 1, 'first'); column2 =
find(horizontalProfile, 1, 'last'); img7 =
img6(row1-1:row2+1, column1-1:column2+1); img7 =
~im2bw(img7); figure; imshow(img7);
title('Detected Number Plate');
imgProp = regionprops(img7,'BoundingBox','Image');
x=1; im0=imread('0.bmp'); im1=imread('1.bmp');
im2=imread('2.bmp'); im3=imread('3.bmp');
im4=imread('4.bmp'); im5=imread('5.bmp');
im6=imread('6.bmp'); im7=imread('7.bmp');
im8=imread('8.bmp'); im9=imread('9.bmp');
imA=imread('A.bmp'); imB=imread('B.bmp');
imC=imread('C.bmp'); imD=imread('D.bmp');
imE=imread('E.bmp'); imF=imread('F.bmp');
imG=imread('G.bmp'); imH=imread('H.bmp');
imI=imread('I.bmp');
imJ=imread('J.bmp');
imK=imread('K.bmp');
imL=imread('L.bmp');
imM=imread('M.bmp');
imN=imread('N.bmp');
imO=imread('O.bmp');
imP=imread('P.bmp');
imQ=imread('Q.bmp');
imR=imread('R.bmp');
imS=imread('S.bmp');
imT=imread('T.bmp');
imU=imread('U.bmp');
imV=imread('V.bmp');
imW=imread('W.bmp');
imX=imread('X.bmp');
imY=imread('Y.bmp');
imZ=imread('Z.bmp'); cnt=0;
out=''; for i = 1 :

5|Page
size(imgProp) this =
imgProp(i).BoundingBox;
if(this(3)>=15 &&this(4)>=30 && this(3)<=80 &&this(4)<=160)
temp = ~imgProp(i).Image; temp=imresize(temp,[40,20]);
imwrite(temp,'temp.bmp'); imgTest =imread('temp.bmp'); img
= imabsdiff(imA,imgTest);
rectangle('position',this,'EdgeColor','r'); minDiff =
sum(img(:)); cnt=cnt+1; ch = 'A';
if cnt==1 || cnt==2 || cnt==6 || cnt==5
img = imabsdiff(imB,imgTest);
if(sum(img(:))<minDiff) minDiff=
sum(img(:)); ch='B'; end img =
imabsdiff(imC,imgTest);
if(sum(img(:))<minDiff) minDiff=
sum(img(:)); ch='C'; end img =
imabsdiff(imD,imgTest);
if(sum(img(:))<minDiff) minDiff=
sum(img(:)); ch='D'; end img =
imabsdiff(imE,imgTest);
if(sum(img(:))<minDiff) minDiff=
sum(img(:)); ch='E'; end
img = imabsdiff(imF,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='F';
end img =
imabsdiff(imG,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='G';
end img =
imabsdiff(imH,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='H';
end img =
imabsdiff(imI,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='I';
end img =
imabsdiff(imJ,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='J';
end img =
imabsdiff(imK,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='K';
end img =
imabsdiff(imL,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='L';
end img =
imabsdiff(imM,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='M';
end img =
imabsdiff(imN,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='N';
end img =
imabsdiff(imO,imgTest);
if(sum(img(:))<minDiff)

6|Page
minDiff= sum(img(:)); ch='O';
end img =
imabsdiff(imP,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='P';
end img =
imabsdiff(imQ,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='Q';
end img =
imabsdiff(imR,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='R';
end img =
imabsdiff(imS,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='S';
end img =
imabsdiff(imT,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='T';
end img =
imabsdiff(imU,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='U';
end img =
imabsdiff(imV,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='V';
end img =
imabsdiff(imW,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='W';
end img =
imabsdiff(imX,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='X';
end img =
imabsdiff(imY,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='Y';
end img =
imabsdiff(imZ,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='Z';
end else img =
imabsdiff(im0,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='0';
end img =
imabsdiff(im1,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='1';
end img =
imabsdiff(im2,imgTest);
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='2';
end img =
imabsdiff(im3,imgTest);

7|Page
if(sum(img(:))<minDiff)
minDiff= sum(img(:)); ch='3';
end img = imabsdiff(im4,imgTest);
if(sum(img(:))<minDiff) minDiff= sum(img(:)); ch='4';
end img = imabsdiff(im5,imgTest);
if(sum(img(:))<minDiff) minDiff= sum(img(:)); ch='5';
end img = imabsdiff(im6,imgTest);
if(sum(img(:))<minDiff) minDiff= sum(img(:)); ch='6';
end img = imabsdiff(im7,imgTest);
if(sum(img(:))<minDiff) minDiff= sum(img(:)); ch='7';
end img = imabsdiff(im8,imgTest);
if(sum(img(:))<minDiff) minDiff= sum(img(:)); ch='8';
end img = imabsdiff(im9,imgTest);
if(sum(img(:))<minDiff) minDiff= sum(img(:)); ch='9';
end end out=strcat(out,ch); end end
msgbox(strcat('Vehicle Registraction Number :',out));

RESULTS:

Input image:

Output :

8|Page
Result Analysis:

For the given respected test images of vehicles , the original vehicle number is
recognized and printed on the screen .The same procedure is tested for multiple
samples of images and the results are verified .

DRAW BACKS:
1) It unable to recognized in blurry and noisy images
2) The data should be manually recorded
3) spaces pictures need to be given as input

Applications:
1. Traffic Monitoring Systems
2. Toll Collection Systems
3. Parking Management Systems
4. Law Enforcement for Traffic Violations
5. Vehicle Access Control in Restricted Areas
6. Automated Vehicle Identification in Highways
7. Stolen Vehicle Detection and Recovery
8. Electronic Road Pricing Systems
9. Border Security and Customs Enforcement
10. Automated Vehicle Registration and Tracking convert
Disadvantages:

1. Environmental Challenges
2. Inaccurate Detection in Challenging Conditions
3. Dependency on Quality of Captured Images
4. Vulnerability to Manipulation and Forgery
5. High Initial Setup Costs
6. Maintenance and Calibration Requirements
7. Legal and Privacy Concerns
8. Limited Performance in Complex Traffic Scenarios
9. Integration Challenges with Existing Systems 10. Potential for System
Downtime and Technical Issues

9|Page
 Conclusion:

In conclusion, the utilization of image processing techniques for automated

vehicle number plate detection presents a promising solution to enhance various
aspects of traffic management, law enforcement, and security systems. Despite
certain challenges and limitations, such as environmental factors and accuracy
issues in adverse conditions, the benefits offered by this technology in terms of
efficiency, accuracy, and scalability are substantial. By addressing these
challenges through continued research and development, automated vehicle
number plate detection systems have the potential to significantly contribute to
improved road safety, traffic flow, and security enforcement in both urban and
rural environments.
REFERENCES:

1. https://www.electronicsforu.com/electronics-projects/hardware-diy/vehicle-
numberplate-detection-using-matlab

2. https://www.mathworks.com/matlabcentral/fileexchange/54456-licence-
platerecognition

3. https://github.com/topics/vehicle-numberplate-recognition?l=matlab

10 | P a g e