Vote Counting Machine (VCM)

Final Year Project Report

Presented

by

Muhammad Taha
CIIT/FA18-BEE-106/ISB

Muhammad Aqeel
CIIT/FA18-BEE-083/ISB

Supervisor
Ahsan Malik
In Partial Fulfillment

of the Requirement for the Degree of

Bachelor of Science in Electrical (Computer) Engineering

DEPARTMENT OF ELECTRICAL ENGINEERING

COMSATS UNIVERSITY ISLAMABAD

COMSATS UNIVERSITY ISLAMABAD

1
 Vote Counting Machine (VCM)

Final Year Project Report

Presented

by

Muhammad Taha
CIIT/FA18-BEE-106/ISB

Muhammad Aqeel
CIIT/FA18-BEE-083/ISB

Supervisor
Ahsan Malik
In Partial Fulfillment

of the Requirement for the Degree of

Bachelor of Science in Electrical (Computer) Engineering

DEPARTMENT OF ELECTRICAL ENGINEERING

COMSATS UNIVERSITY ISLAMABAD

COMSATS UNIVERSITY ISLAMABAD

2
 Final Approval
This Project Titled
Vote Counting Machine
Submitted for the Degree of
Bachelor of Science in Electrical (Computer) Engineering

by
Name Registration Number

Muhammad Aqeel CIIT/FA18-BEE-083/ISB

Muhammad Taha CIIT/FA18-BEE-106/ISB

has been approved for

COMSATS UNIVERSITY ISLAMABAD

_____________________
Supervisor
Name,
Designation

______________________ ______________________
Internal Examiner-1 Internal Examiner-2
Name, Name,
Designation Designation

_____________________
External Examiner
Name,

Designation_____________________

Head
Department of Electrical and Computer Engineering

3
 Vote Counting Machine
An Undergraduate Final Year Project Report submitted to the
Department of
ELECTRICAL AND COMPUTER ENGINEERING

As a Partial Fulfillment for the award of Degree

Bachelor of Science in Electrical(Computer) Engineering

by

Name Registration Number

Muhammad Aqeel CIIT/FA18-BEE-083/ISB

Muhammad Taha CIIT/FA18-BEE-106/ISB

Supervised by

Ahsan Malik
Professor,
Department of Electrical Engineering

CU Islamabad

COMSATS UNIVERSITY SLAMABAD

Feb 2021

4
We dedicate this small effort to our parents and teachers who have been a source of inspiration
for us. Without their utmost support, this was difficult to achieve. We thank them from the core
of our hearts.

5
 Declaration

We, hereby declare that this project neither as a whole nor as a part there of has been copied out
from any source. It is further declared that we have developed this project and the accompanied
report entirely based on our personal efforts made under the sincere guidance of our supervisor.
No portion of the work presented in this report has been submitted in the support of any other
degree or qualification of this or any other University or Institute of learning, if find we shall
stand responsible.

Signature:
Name: Muhammad Taha

Signature:
Name: Muhammad Aqeel

6
 Acknowledgements

Primarily, we are grateful for Almighty's ability to help humanity develop. Allah is the one who
made universe and blessed us with our beloved parents and friends who contributed towards our
studies. This project would not have been possible without their direction and support.

We would like to begin saying thanks to Sir Ahsan Malik, our honorable project advisor, who
helped us a lot in fulfillment of this project. It was his divergent thinking which led us to this
point and provided us with this excellent opportunity and encouraged us at every bold step and
gave us his productive feedback in our whole project journey. It is a privilege to work under him.
A special thanks to Sir Muhammad Mehr Nawaz for helping us at every single turn and obstacle
that we faced along this journey of ours.

7
Table of Contents
Abstract...................................................................................................................................................... 13

1.Introduction ............................................................................................................................................ 14
1.1 Overview ............................................................................................................................................... 14
1.2 Generic electronic voting phase………………………………………………...…………………… 14
1.3 International status of electronic voting……………………………………………...……………… 16
1.4 Comparing Features…………………………………………………………………….…………..…17
1.5 Comparative Analysis with electronic voting system……..…………………………………..………19
1.6 Problem Statement………………………………………..…………………………………..………20

2.Literature Review……...………………………………………………………………………………22
2.1 Image Processing…………..………………………………………………………………………….22
2.2 How digital image is form…….………………………………………………………………………24
2.3 OpenCV……………………………………………………………………………………………….25
2.4 Removing Background Noise Using OpenCV………….…………………………………………….25
2.5 Template Matching…………….…………………………..………………………………………….25

3.Design (H/W) ………………………………………………...………………………………………...33

3.1 IR Sensor……………………………………………………….……………………………………...34
3.2 775 Motor…………………………………………………………..………………………………….35
3.3 Solenoid……………………….………………………………………………………………………36
3.4 Arduino Uno……………………………………………………………….………………………….37
3.5 Bug Converter………………………………………………………………….……………………...38
3.6 DC Voltage Regulator……………………………………………………………...………………….39
3.7 MOSFET………………………………………………………………………………………………39
3.8 Raspberry Pi 4……………………………………………………………………………...………….41
3.9 Pi Cam………………………………………………………………………………………...…….…43

8
4.Design (S/W) ………………….………………………………………………………….…………….49
4.1 Arduino………………………………………………………………………………………….…….49
4.2 Proteus………………………………………………………………………………………..….…….50
4.3 Simulation on Proteus…………………………………………………………………………………50
4.4 PCB Design……………….………………………………………………………………….………..51
4.5 PyCharm…………………….………………………………………………………………………....51
4.6 OpenCV……………………………………………………………………………………………….53
4.7 Flow Chart…………………………………………………………………………………………….53

5.Mechanical Design ………..….………………………………………………………….…………….54

5.1.OMR……………………………………………………………,,,……………………………………54
5.2. Limitation……………………………………………………..………………………………………55
5.3.Design…………………………………………………………………………………………………55

Conclusion………………………………………………………………………….…………………….59

Expected Outcome…………..………………………………………………………..………………….59

References……………………………………..…………………………………………….……………60

9
 List of Acronyms
SSD ………………………………………………………Sum of Squared Differences
NCC …………………………………………………….Normalized Cross Correlation
GPIO……………………………………………………. general purpose input/output
EVM………………………………………………………Electronic Voting Machine
APIs…………………………………………………………. Application Programming Interface

10
 List of Figures
2.1 steps involved in processing of an image……………………………………………………..…22
2.2 different phases in digital image processing…………………………………………………..…23
2.3 weights are added in image by the system……………………………………………………..…25
2.4 steps involved in background removal of an image..…………………………………………..…25
2.5 Noise removing……………………………….………………………………………………..…26
2.6 ballot paper with noise……………………………………..…………………………………..…27
2.7 ballot paper without noise…………………………………………….………………………..…27
2.8 how weights are added in image...……………………………………………………………..…28
2.9 flowchart of removing background of an image……...………………………………………..…30
2.10 template matching……………………………… ……………………………………………..…31
2.11 refrence image……...…………………………………………………………………………..…31
2.12 template………………………………………….……………………………………………..…32

3.1 block diagram…………...………………………….…………………….………………………33

3.2 PCB hardware……………...…………………………………………….………………………34
3.3 IR sensor………….………………...….…………………………………..…………………….35
3.4 775 motor ………………......….……………………………………………..…………….……36
3.5 solenoid………………......….…………………………………………………..……….….……36
3.6 Arduino…..…………......….…………………………………………………………....….……37
3.7 GPIO……....……………......….…………………………………………………………………37
3.8 buck converter…………......….…………………………………………………………….……38
3.9 Voltage regulator….…….....….…………………………………………………………….……39

11
3.10 IRFZ44N….…………….....….…………………………………………………………….……40
3.11 raspberry pi 4b….……………..…………………………………………………………….……42
3.12 GPIO………….. .……………..…………………………………………………………….……42
3.13 pi cam…….…………………....…………………………………………………………….……43
3.14 Data sheet………….………………..……………………………………………………….……43
3.15 Flow chart………….………..………………...…………………………………………….……45
3.16 usb communication………….…………..………………………………………………….……46
3.17 communication through Rx and Tx……………...………….……………………..……….……46
3.18 PCB design………………………………….……………..……………………….……….……47
3.19 resistors……………………………………...………….…………..……………………….……47
3.20 LEDs……………………………………….…………..…………………………..……….……48

4.1 Arduino IDE……………………………………….…………..…………………………...……49

4.2 Proteus simulation…………………………………….…………..………………………...……50
4.3 PCB ……………………………………….………….…………..………………………...……51
4.4 Pycharm …………….………………….……….…………..……………………………...……52
4.5 Flowchart…………….…………………….………………………………………………..……53

5.1 OMR….……………......….…………………………………………………………….….….…54
5.2 Acrylic Sheets…………………………………….……………...……....….……………..…….56
5.3 Mechanical Design ……………………………….……………...……....….……………..…….56
5.4 Mechanical Design ……………………………….……………...……....….……………..…….57
5.5 Iron rods………………………………………….……………...……....….……………..…….57
5.6 3d printed cap…………………………………….……………...……....….……………..…….58
5.7 3d printed cap…………………………………….……………...……....….……………..…….58

12
 List of Tables

1.1 COMPARATIVE ANALYSIS WITH ELECTRONIC VOTING SYSTEMS…………,,,…………….19

13
 Abstract
Vote counting is a time-consuming task. Not only counting votes is slow and labor intensive, but
it is also insecure because the votes are prone to both intentional and inadvertent human error.
While there are EVM solutions available around the world that solve this issue, they need a
change in the entire election process, and gaining trust is a separate issue. Our solution comes
into play here. We're working on a quick and efficient vote counting mechanism based on image
processing.

Chapter 1

Introduction

1.1 Overview
14
The current voting system of Pakistan is unreliable. The results are not accepted by the opponent
parties and often turns into agitations and due to which our economy/financial system is affected.
Furthermore, t s costly and create burden on our economy. Results can be tampered by the voting
staff. Hence present government decided to start E-voting system in our country which is
comparatively beneficial and reliable.
Electronic voting is used to make the voting process easier and more efficient. Proper
management will improve the security of ballots and speed up ballot counting. If the entire
election process is not carefully planned or the EVM is not carefully designed, it can undermine
the credibility of the entire election process.

1.2 Generic Electronic Voting Phases/Stages.

There exists a wide variety of voting/electoral processes. Electoral process is everything that has
to do with holding an election. The end-to-end voting process is divided into several phases or
stages which are identified as follows:
1. Registration:
This phase involves compiling a list of people eligible to vote. In addition, the ballots are
prepared and all the arrangements or the activities needed for the election.
2. Validation:
Involves verifying the details of the voters and only allowing the registered voters who
have not already voted to proceed.
3. Collection:
Involves collecting the voted ballots from all over the country.
4. Verification:
Involves the verification of ballots. Only the valid ballots are used in the next phase.
5. Tallying:
This phase involves in counting the votes. After this, the results are published.
6. Claiming:
All the claims (if any) are investigated.

Benefits of E-Voting

 Speed up the process of counting votes

 It is more efficient than our current voting system
 Improve the presentation of the ballot papers
15
  It makes easier to cast the vote
 Less spoilt ballot papers
 It saves the cost of elections in long term
 It can increase the number of voters due to its global reach

Weakness of E-voting

 It will reduce the influence of election administration

 High security is needed to protect EVM during and after elections
 Maintenance cost and purchasing cost
 Risk of manipulation by hackers or any insider
 Lack of public trust
 It will limit the recount
Since there is not such a thing as perfect, so it is necessary to choose the right system by
carefully weighting the advantages and disadvantages.

Types of E-Voting systems

Direct recording electronic voting machine (DRE)

They have been in use since 1990s in the USA. In this system voter cast the vote directly via
computer screen or touch pad. After the election process machine printed the results of election.

OMR System
When the voters cast their votes the ballot paper is directly feed in the scanner. Scanner scans the
ballot paper and then count the vote. Our system is also works on this principle.

Electronic ballot printers

It is almost same as DRE the machine printed the token that is feed in scanner and vote is casted.

Internet Voting System

This system uses the internet to cast the vote. Vote can be cast by computers or any devices those
have internet connection. It is the easiest way to cast the vote, you don’t need to go anywhere to
cast the vote.

16
Controlled and Un-controlled electronic voting
Controlled Electronic Voting happen when voting is taking place in polling station under a
supervision. DRE and OMR system are example of it.
Uncontrolled Electronic Voting happens when vote is cast without any supervision. Internet
Voting System is example of it. In Uncontrolled Voting System there is concerns of vote buying,
identity crises etc.

1.3 ELECTRONIC VOTING INTERNATIONAL STATUS

Brazil

The electronic voting equipment used in Brazil is unique in that it counts the votes after they are
cast, producing both digital and printed records of the number of votes cast for each candidate.
12,000 machines used to produce a paper ballot that the voter could peruse and deposit in a box
for recount. These paper-trail machines were used during the election.

India

In India, the first electronic voting election will take place from April 20th to May 10th, 2004.
With more than a billion inhabitants, India is the largest democracy in the world. With a
population of approximately 668 million and more than 1 million EVMs need to be deployed.
EVM has been used in many state elections since its legalization in 1989, but not in federal
elections. EVM is being developed by Electronics Corporation and Bharat Electronics of India.
EVM consists of two parts. One is managed by the voting staff and the other is for voters. The
control unit records the vote when the voter presses the button next to the voting unit candidate's
name and icon. A vote was taken when the light next to the button turned on and a short beep
was heard. The voter then presses a button to reset the machine for the next voter. In rural areas,
the EVM comes with a carrying case and can battery powered. Election Commission officials
claim that each EVM can get 5 votes per minute, or 3,000 votes per election day.

Italy

Italy's electronic voting system was involved in the 2004 general election. According to the
Italian government, the main advantages of the electronic voting system are a simpler and faster
process, more accurate ballot counting, faster and more secure transmission of results, and

17
improved overall voting efficiency. The technical specifications of the electronic ballot counting
system have not yet been published by the Italian government. A special national committee
evaluates pilot projects in terms of system efficiency and addresses potential issues. The
Commission then makes the recommendations needed to prepare the system for more extensive
testing in future elections.

Spain

Spain has attempted to vote on diverse types of computers. In the March 14, 2004, general
election, several small, non-binding electronic voting trials were successfully conducted. In
addition to computer-based voting systems, primarily based on Irish models, these included a
variety of technologies, including internet and SMS remote voting. In the Catalan parliamentary
elections on November 16, 2003, three electronic voting pilot tests were successfully conducted.
For living abroad, this includes remote voting over the Internet and touch screen voting using an
electronic counting system.

1.4 FEATURES TO COMPARE

This section refers to some of the key aspects of the evolution of electronic voting and
provides a comparison of the characteristics of different systems. You can find these three
aspects of the national system investigated

A. Whether the national system uses a paper trail.

B. Whether system allows unidentified blanks.
C. Whether software is available publicly.

D. Paper audit trails:

Only Brazil, Of the 10 countries surveyed, we used extensively paper audit reports. The
paper audit trail will be phased out in Brazil by October 2004. For the October 2002
elections, the Brazilian government deployed them on a limited scale, with paper audit trails
being used on 12% of all machines This system allowed voters to see the wording of a vote
before both paper and electronic votes were recorded and stored. The paper audit trail will
be phased out in Brazil by October 2004.

E. Basis by which system was introduced

18
 Electronic voting was first used in limited constituencies or local elections in all 10 countries
surveyed. Ireland, which was the first country to use electronic voting in three constituencies
in the 2002 election, also falls into this group. In addition, testing is underway in many
countries where national elections are not yet required to conduct full electronic voting.
Authorities in various countries (such as Brazil and Australia) checked the results of the
selected machines to make sure they were correct.

F. Treatment of blank or invalid votes

Brazil and Australia are only countries those can throw blank ballots and both keeping
voters identities secret. In Brazil, blank votes are counted as part of the total ballot, but in
Australia they are not. The Indian system does not allow invalid voting for a variety of
reasons, including the fact that the number of invalid votes has historically been relatively
high and one of the main advantages of electronic voting is the ability to reduce the number
of incorrect votes. Blank or incorrect ballot papers are no longer available in the Belgian
election system.

G. Open source versus proprietary software

In two of the countries surveyed (Australia and Belgium), the source code of electronic
voting software is available on the Internet. Australia was the first country to publish its
software source code to increase public confidence in the system, followed by Belgium in
1999. For a limited time before elections, Brazil allows the partial insight into the source
code. It is only available to political parties and election committees. The software source
code is not publicly available in India.

1.5 COMPARATIVE ANALYSIS WITH ELECTRONIC VOTING

SYSTEMS
Coun E- Compa Electio Electoral Introd Year Software Hardwar Problems
try Voting ny n Type System uced Used Used e Used
Year

19
India 668 BHEL State FPP 2001 2009 EPROM EVM None
million

Belgi 3.2 Steria Genera Open PR- 1994 1999 Digivote, DEVS 2003: 500 power
um million l List Jites, and
& Stesud computer failure
Munici
pal
Brazi Sixty- UniSys All 1996 1996 GEMS GX-1 None
l six & Govt. integrated
million Diebold Level processor
Austr 218000 Softwar ACT PR-STV 2001 2001 eVACS PCs None
alia e federal
Improve
UK 1.5 SVS Local FPP 2000 2000 AVC DRE Mobile e voting
million Govt /2003
Spain 3000 Indra Munici PR-List 2002 2003 SIRE SIRE None
pal System
Cana 98000 CanVot Munici FPP 2002 2003 CanVote CanVote None
da e pal on Linux Internet
Table 1.1

E-VOTING DEVELOPMENT PHASES Mechanical and electronic

voting initiative: 1849– 1900

In 1856, the Australian state of Victoria became the first to use standardized official ballots. On
this sort of paper ballot, all candidates and issues are listed in a fixed sequence and tabulated by
hand. n 1888, Massachusetts became the first state in the United States to implement the
Australian ballot system statewide. Herman Hollerith receives a patent in January 1989 for a
method of generating data for the US Census using punched cards. Later in November,
Rochester, New York`s Jacob H. Myers has patented the Myers automated booth, the first
mechanical lever voting mechanism. This eliminates overvoting, speeds up the ballot counting
process, and significantly reduces the risk of unjustified voting counts, as votes are counted
mechanically. This machine firstly used in New York in 1892.
1849 De Burettes develops an electronic decision-making telegraph. 1856 Australia first place to
use uniform official ballots. 1859 “Werner von Siemens” Take this idea further with your first
application. 1865 The first automated decision-making (Yes/No decisions) method used. 1869
Thomas Edison developed and patented electro-mechanical device. 1888 Massachusetts adopts
Australian secret ballot. 1889 Punch Card System Patented Mechanical Lever Machine Patented
1892 Lever Machine first used n Lockport, New York. 1895 The first electronic voting
company was established based on the invention of Jacob H. Meyr. 1900 Japan started first
secret ballot voting. On Dec. 14, 1900, the U.S. standard voting machine company was formed.

20
1913- 14 Allan Walsh of new jersey introduced an electrical and mechanical system of vote
casting. 1915- 16 installed electrical voting system in the Wisconsin legislature, by
Representative William Howard of Georgia. 1930 Used Lever Machines in Every Major US
City. 1939 The reorganization act enacted April 3, 1939. 1940- 1950 Don A. Allen support
adoption of voting machines for Los Angeles. 1955 "Erich Fromm" presents ideas for
communication and decision making through networked technology devices. 1960 The first
computers for tabulating votes have been developed. The first punch card machines have been
developed and implemented. 1962 First Optical Scan Ballots used in Kern City, California.
1965Punch Card Voting System Patented. 1970 “Murry Turoff” developed a computer supported
Delphi panel. 1974 Direct Recording Electronic (DRE) Voting Machine Patented. 1977 Precinct-
Based Optical Scan System Patented. 1982 Nebraska First to Officially Use American
Information Systems (AIS) Central-Count Ballot Tabulator. 1987 Shouptronic Electronic Voting
Machine Patented. 1988 Report Warns of Problems with Pre-Scored Punch Cards. 2000 The
University of Osnabrück develops and implements an online voting system. The CANN elects its
five directors via the internet. 2002 OCT. 29 The first e-Voting election in Japan at Niimi city of
Okayama prefecture. This election ended successfully. Georgia s the first company in the
country to use a direct recording electronic voting machine. India successfully used EVM n the
entire country. 2003 DEC. 9 Election Systems Companies Form Information Technology
Association of America (ITAA). 2004 Nevada Mandates Voter-Verified Paper Audit (VVPA).

1.6 Problem Statement:

Our current voting system s unreliable and have many disadvantages. The results are not
accepted by the opponent parties and often turns into agitations and due to which our
economy/financial system is severely affected. Furthermore, t s costly and create burden on our
economy. The voting staff can tamper results. Hence present government decided to start E-
voting system in our country which s comparatively beneficial and dependable. The advantages
of EVM are as follows
1. Quick/ Fast results
2. Less Expenditures
3. Error Free
4. Results cannot be tampered easily
5. Transparent Voting system
6. Results are less Disputed
7. Less Manpower required

21
While these EVM solutions exists, they require change of whole election process and building
the trust of people s another problem. That s where our solution comes into play. We are
developing a vote counting machine based on image processing that will be fast and efficient.
Our aim is to make vote counting more efficient and secure and we are trying to achieve this
without changing the whole Election process, so that front-end of the process remains same but
back-end s modified to be more efficient.

Chapter 2

Literature Review

22
2.1 Image processing

Image processing is a technique that converts an image into a digital format and then performs
operations on it to create a better image or extract valuable information. This is a signal
distribution method where the input is an image similar to the following. It can be a video frame
or a photo, and the output is an image or image-related feature. In most image processing
systems, images are treated as two-dimensional signals and signal processing algorithms are
applied.

It has numerous applications in various fields of engineering and computer science and is one of
the fastest growing technologies today, and image processing is an important research topic.

Image processing includes the following three steps:

1) Using a scanner or digital image to import the image in device.

2) Analyzing and changing the image, such as reduce size and image orientation, as well
as recognizing patterns that are not visible to the naked eye.

3) The final stage is output, which can be a transformed image, or a report based on
image analysis.

Figure 2.1 (Steps involving in processing of an image)

Purpose of Image processing

The purpose of image processing is divided into 5 groups. They are:

1. Visualization – Observe the objects that are not visible.

2. Restoration of image – To create a better image.

3. Image retrieval – Seek for the image of interest.

23
4. Measurement of pattern – Measures various objects in an image.

5. Image Recognition – Distinguish the objects in an image.

Types

There are two types of image processing methods, analog and digital. Analog or visual image
processing methods can be used for tangible copies such as printouts and photographs. Image
analysts use a variety of interpretation principles when applying these visual techniques. Image
processing is limited to the knowledge of the analyst as well as the area to be analyzed. Another
important element of image processing through the visual process is association. As a result,
analysts use a combination of personal expertise and secondary data to interpret the image.

Use your computer to edit digital photos using digital processing techniques. The raw data from
the image sensor on the satellite platform is incorrect. Many steps need to be taken to overcome
these vulnerabilities and make the information clearer. When using the digital approach, all types
of data must go through three general phases: preprocessing, expansion and display, and
information extraction.

Different Phases in Digital Image Processing (fig 2.2)

There are different phases in image processing first is preprocessing, second is applying weights
to image and train the data, third is decision and classification of data, fourth is checking the
output accuracy and then there is an output.

24
2.2 How a digital image is formed

To form a digital image, we use sensor array. When light shines on an object sensor will detect
the amount of light that is reflected by the object. This data from these sensors is in form of
continuous voltage signal to form a digital image it is converted into digital form. After
converting into digital form, a two dimensional array matrix is formed which is then saved as a
digital image.

Machine/Computer vision

The development of a system in which the input a image, and the output is some type
information is known as machine vision or computer vision. Consider developing a technology
that scans a person's face and unlocks any lock. This is what the system would look like.

Figure 2.3(Weights are added in image by the system)

2.3 Open CV

The image processing package open cv in Python is used in this project. OpenCV is an excellent
tool for image processing. It is an open-source software for tasks including facial recognition,
object tracking, and landmark detection, among others. Python, Java, and C++ are among the
languages supported. However, for the purposes of this post, we shall exclusively discuss
Python.
The library gives you access to hundreds of useful functions and algorithms, all of which are free
to use. Some of these features are common and appear in computer vision work.

25
2.4 Removing Background Noise using OpenCV

Background noise is prevalent while working with photos. This OpenCV function allows you to
extract only the information you need from an image while discarding the rest.

Figure 2.4(Steps involving in background removal of an image)

The first set contains four dots that cover the foreground area. If you look closely, there are four
points in the image above. The first set consists of the coordinates of these points. The
boundaries of the original image are different points. Then use the cv2.getPerspective () function
to get the output matrix and use the cv2.warpPerspective () function to convert it to the required
format.

Figure2.5(Noise Removing)

26
In this image the background noise is removed using python code. First ewe selects four points
around the image, then we form the square around the image and then remove the noise. This
algorithm is used in processing of ballot paper.

Figure 2.6(ballot paper with noise)

In this ballot paper we are taking four points across the corner of the black line. Then algorithm
makes the line across it and then remove the noise in this image. Image after noise removal look
like this

Figure 2.7(Ballot paper without noise)

27
First the image is converted into gray scale, then the edge detection is applied, and the contour of
the image is found. Any contours that is big or small then the foreground is removed. Then the
mask is generated from contours after this it is blend into the output image. This process can also
be done by using artificial neural network or deep learning, but it is very complex method.

Figure 2.8(How weights are added in a image)

This flow chart shows the process that involves during the background removal of an image.
First, we import the image, then apply different weights on after this we measure accuracy and
then our final result is ready.

2.5 Fast Template Matching with Polynomials

Algebraic template matching quickly calculates the similarity between a template and a sub-
image of an input image of various widths and heights, given the template and the input image.
For any location, width, and height, the partial image most like the template image is recognized
from the input image. Instead of using the template picture, the suggested technique uses a
polynomial that approximates it to match the input image. The suggested approach works well,
especially when the template picture's width and height deviate from the partial image to be
matched. For efficient a rough representation of the template picture, an approach based on the
Legendre polynomial is presented. This approach increases the quality of the approximated
image while reducing processing expenses. The computational cost of the suggested technique is
proven to be lower than that of existing methods, both theoretically and experimentally.

28
Template matching using the square of the difference and the sum of the
normalized cross-correlation
In digital image processing, template matching is a way of detecting the location of a sub picture
within an image. A part of the image is commonly compared to the template, which is a sub
image. The template could come in a variety of sizes, colors, and shapes. Template matching is
required for both image registration and object recognition. This study examines the performance
of the Sum of Squared Differences (SSD) and Normalized Cross Correlation (NCC) image
registration methods in matching a template to a picture. The purpose of this experiment is to
compare the capabilities of both strategies in terms of output quality and execution time. It also
evaluates the effect of the template image on the output image in the presence of noise and
rotation.

An Overview of various template matching methods in image

processing
The computer vision community believes that the detection and classification of objects in
photographs has increased significantly. If you are uncertain about the scale and rotation of the
actual target, a common image processing problem is to use a template to determine the
proximity of the object. Templates are subparts of elements used to match completely different
things. The versatility and ease of use of the template matching approach makes it one of the
most used object localization techniques. Image processing, signal and video compression, and
pattern recognition are examples of areas where template matching is used. This white paper
outlines applications and approaches that use template matching techniques.

Evaluation of Voting with Form Dropout Method for Ballot Vote

Counting

The accuracy of ballot counting has become a well-known concern in the voting process. Digital
image processing techniques can be used to inspect printed ballots. The expected audio trajectory
depends on the current image processing algorithm of the audio acquisition procedure. These
methods do not respect annotations created outside the requested input field. You can take any
mark on the form and use it on your machine and use different approaches to exit the form to
draw informed conclusions. Even with best practices, some markers are missing, resulting in
false positives. This study explores how detection can be improved by adjusting the results of
multiple tag extraction algorithms. Added versatility with simple graphics.

Template matching is a process in which we provide different image to the algorithm and then
algorithm compare the images with the input image and give us similar output. Template

29
matching work as matching each pixel of input images and give the result of array similar values
of template image.

Figure 2.9(Flow chart of removing background of an image)

The template image slide over the input image then it compares the each pixel of image after this
a threshold is set if the result is greater than the threshold the output will be provided otherwise
some weights are added on the image and again threshold is applied.

30
 Figure 2.10(Template matching)

In this image there is a template image of the bird and input image of the sky with birds. The
template image simply slides over the input image and compare each pixel after which threshold
is applied and we get the output image of the bird.

Figure 2.11(Reference image)

In this image, the image of the stamp is provided as the template image and the whole image as
input image. The image of kite, cycle, tiger, arrow also provided to algorithm with the label on it.
The template image slides over the input image then compare the each pixel and threshold is set.
At which point the template is matched the vote is cast to that party represented by each image.

31
 Figure 2.12(template)

This image shows that the template is matched on the place of bat, so vote is casted to the bat.

32
 Chapter 3
Design (H/W)

Block Diagram

Figure 3.1(Block Diagram)

The power supply is providing 24v DC to the buck converter. Then 12 v is supplied to another
buck convertor which reduce the voltage to 5v which is provided to Arduino and Raspberry Pi.
Camera is connected to Raspberry Pi and IR sensor, MOSFET, Motor, Solenoid are connected
with the Arduino. Arduino is connected serially via USB with Raspberry Pi.
The input is taken from camera and IR sensors connected with Raspberry Pi and Arduino. Output
is given to motor and solenoid connected to the Arduino. The end result of voting process
uploaded to the server.

33
PCB

Figure 3.2 (PCB Hardware)

PCB is the printed circuit board which contains the tracks made from copper or other conducting
material. It eliminates the use of wires and give a look too the circuit. It is basically made of a
nonconductive material like fiber or plastic etc. with a sheet of copper on it. There are different
types of PCB like single layer, multi-layer, double sided, flex, rigid PCB. We are using single
sided PCB for our project. It is easy to make and east to sold. The other types of PCB are use in
industries.
Our PCB contains MOSFET, resistors, buck converter, relay, motor driver ic, t-blocks and
Arduino. We are providing 24v DC to the buck converter which converts voltage in 12 and 5
volts. 5 volts are used to power the Arduino and 12v is used to power the motor. Direct 24v from
supply are used to power the solenoid.

3.1 IR sensor

This IR proximity sensor is a versatile infrared sensor that can be used for obstacle detection,
color detection, fire detection, line detection, encoder detection and more. The sensor provides a
digital output.

34
If there is a black line in front of the sensor, the sensor will output Logic 1 (+ 5V) as a digital
output and the white line in front of the sensor will output Logic 0 (0V). An integrated LED
indicates the presence of paper. Arduino uno connects directly to this digital output.

Figure i3.3(IR sensor)

IR sensors are helpful in close range. They work on infrared radiation. They emit the Infrared
radiation and then the light is reflected, they detect the light and provides the input to the micro
controller. IR sensor collects the maximum light that reflect to give maximum accuracy.
This sensor is used to take the input that weather the paper is loaded in machine or not, they are
also used to calculate numbers of ballot papers loaded inside the machine.

3.2 775 Motor

The size ID of a motor recognized as a standard is 775. It is a high-torque, high-power gadget

designed for robots, quadcopters, and industrial applications.
In this project motor is coupled with two rollers that will load the papers in and out of the paper
tray.

Fig3.4(775 Motor)

35
Motor is use for roller. When the roller moves ballot paper will comes out of the machine after
getting signal from raspberry pi.

3.3 Solenoid

Solenoid switches are electric switches commonly used when low current switches are used to
activate high current circuits such as electric motors.
Turn the key switch to start and when the transmission is in neutral, the circuit between the
battery and solenoid closes.
A heavy-duty switch connects the vehicle's battery to the starter motor when the solenoid switch
is activated.

Fig 35(Solenoid)

The solenoid in used in this project to load the ballot paper in the machine so that the camera
takes the image of ballot paper and roller starts to move.

3.4 Arduino Uno

The Arduino UNO is an inexpensive, customizable, and easy-to-use open-source programmable

microcontroller board for a variety of electronics projects. Arduino UNO can connect with other
Arduino boards, Arduino shields, and Raspberry Pi to power relays, LED, sensors, and motors.

36
 Fig 3.6(Arduino)

Fig 3.7(GPIO)

GPIO are the general input and output pins. Arduino has 22 GPIO. There are 5 ADC pins those
are use to convert input analog data into digital data. It has atmega328p controller which works
on 8MHz frequency.
Arduino Uno is cheap micro controller and provide large numbers of GPIO pins, it work on
8Mhz oscillator frequency and have AT Mega 328p micro controller in it, so it is fast enough to
process our hardware in minimum time.

37
3.5 DC To Dc Buck Converter Step Down Module Power Supply

The LM2596 step-down DCDC step-down converter power supply is a step-down (step-down)
switching regulator. LM2596 Buck can drive a 3-A load with excellent line and load regulation.
These devices are available in fixed output voltages of 3.3 V, 5 V, 12 V, and an adjustable output
version. The LM2596 buck module operates at a switching frequency of 150kHz, allowing it to
use fewer filter components than low frequency switching regulators.
It consists of a diode that is use for switching. Switching helps to step down the voltage.

Features Of LM2596 DC-DC Adjustable Buck Converter Step

Down Module:

Non-isolated step-down (BUCK) switching regulator.

Short circuit protection: Current limiting, self-recovery
Direction of potentiometer setting
Asynchronous correction

Fig 3.8(Buck Convertor)

3.6 DC voltage regulator

The voltage regulator provides outstanding load regulation and linearity regulation, as well as
many protections, including over-current, over-temperature, and short circuit protection.
Voltage regulating mode: PWM modulation

38
Input voltage: DC 4-38V
Output voltage: DC 1.25-36V
Maximum current: 8A

Fig 3.9(Voltage Regulator)

This is the buck convertor that require 24v input and we can get desire output by rotating the
potentiometer. It works on the switching mechanism to step down the voltage, it uses diode for
this purpose. The input is DC and output is also DC. It provides output less than 24v. It also
contains an LC filter which provide us stable output current. Transformer are used to step down
the voltage, but they work on AC, for DC we use buck converter.

3.7 MOSFET

IRFZ44N is a common MOSFET transistor that can be used in a wide range of general-purpose
applications. The transistor has high-speed switching capabilities, making it excellent for use in
applications where high-speed switching is required.

39
 Fig 3.10(Z44N)

IRFZ44N Features

 Advanced Process Technology.

 Ultra-Low On-Resistance.
 Dynamic dv/dt Rating.
 175°C Operating Temperature.
 Fast Switching.
 Fully Avalanche Rated.
 Lead-Free.

SYMBOL PARAMETER MAX. UNIT

VDS
Drain-source voltage 55 V
ID Drain current (DC) 49 A
Ptot Total power dissipation 110 W
Tj Junction temperature 175 ˚C
RDS(ON) Drain-source on-state 22 mΩ
resistance VGS = 10 V

tab
d

1 i2 i3 s

IRLZ44n MOSFET has an internal diode that helps to control the back emf that is generated
while using the motor. Without this the MOSFET will burnout and it can also damage the micro
controller as well.

40
It ha 3 pins first is source, second is gate and third is drain. First in connected with the micro
controller for switching, second is connected with positive terminal of motor and third is
connected with the ground, the ground of controller and ground of motor must be common.
This is widely use MOSFET. We are using this for switching the motor and for controlling the
PWM (speed) if the motor. It has high current rating so there is less chance of this MOSFET to
burn.

3.8 Raspberry pi 4

Raspberry pi is a Linux based microprocessor unit that is used for various uses such as image
processing, embedded systems, controlling electronic circuits and IOT. It has Dual core
processor.

Raspberry Pi 4b Tech Specs

Quad core Cortex-A72 (ARM v8) 64-bit

2GB, 4GB or 8GB LPDDR4-3200 SDRAM (depending on model)
2.4 GHz and 5.0 GHz IEEE 802.11ac wireless, Bluetooth 5.0, BLE.
Gigabit Ethernet.
2 USB 3.0 ports; 2 USB 2.0 ports.

Fig 3.11(Raspberry Pi 4b)

Software

41
It works on Raspbian OS 64bit which is provided by raspberry pi. It can also work on Windows
or Linux. It can also be connected remotely with desktops or PC. It is recommended to use
Raspbian OS.

GPIO Connector Pinout

Figure 3.12(GPIO)

GPIO are general input and output pins that’s allows to interface different sensors, motors with
the micro controller. Raspberry pi 4b ha 40 GPIO that allows user to use wide range of
components and not to worry about input and output pins. It requires 5v input to operate and also
provides 5v at GPIO.

3.9 Pi camera

The Pi Camera Module is a small, lightweight camera that runs on the Raspberry Pi. The MIPI
camera serial port protocol is used to communicate with Pi. Image processing, machine learning,
and surveillance programmers are common applications.
We are using the 8 mega pixel camera module of pi cam which allows to capture the image in
high resolution, and it has a ultra-wide camera that helps us to capture the image in closer range.

42
 Fig 3.13(Pi Cam)

Specification

Figure 3.14(Data Sheet)

The main feature of this camera is that it makes focus automatically, there is no need to adjust
the focus manually. It can operate on just 5v and provide clear output images without noise. It
also comes with night vision lens which can be attached on it so it can provide images at night,
but the images are not enough clear so that they can be processed, there is a lot of noise and les
data in the image.

DC Power Supply

This project uses 24-volt power supply with 1.5 A of current. This supply converts AC into DC
and provides 24v in output. The output is pure DC.

Methodology:

This project uses Arduino Uno as a microcontroller based controlling unit which relates to 2 IR
sensors that are used as input to detect the presence of paper. A 775 motor and a solenoid will
function as an output device that will take the input from sensors and then move accordingly.

43
Motor is coupled with two rollers that will load the papers in and out of the paper tray. Up and
down motion of solenoid is used to load the papers into rollers. When solenoid is dropped it will
load the uppermost paper into the roller and 1st IR sensor will detect the paper and the motor will
start. When the paper reaches the 2nd IR sensor the solenoid will go up and only second roller
will move the paper. Motor will stop moving when both sensors did not detect the paper.
For the counting and image processing, microprocessor Raspberry pi 4B 4GB variant is used.
We use different microprocessor for this purpose because the image processing must be fast for
this system to be efficient. Firstly, the Pi Cam will capture the snapshot of the ballot paper and
then covert it into Grayscale image. Then we will use Sum of squared difference-based template
matching to detect the voted candidate on that ballot paper. Next, the number of votes for this
candidate will increase. A threshold must also be taken into the account so that the number of
inappropriate votes can be reduced.
Finally, an appropriate delay will be added between the loading of ballot paper in roller and
starting of the motor. This delay must be equal to the time it takes for the camera to take
snapshot and process the image. This delay is added to make sure that the image processing and
loading and unloading of paper are synchronous.
The Arduino will start to work when it gets serially input from raspberry pi. When image is taken
by the camera raspberry pi will give the signal to Arduino using USB port so the Arduino will
start its work. The motor will start to rotate which rotates the roller. When roller moves the paper
will come out from the machine and then solenoid switches which loads the next paper in the
machine.

Flowchart:

44
 Fig 3.15(Flow Chart)

First, we take the input image of ballot paper from camera which is attached with raspberry pi.
When camera take the image of ballot paper the roller will start to move, and the ballot paper
move out. Then the camera take will the second image, this process will continue in loop until no
ballot paper left inside the machine.

Raspberry Pi and Arduino Communication

It is basically an asynchronous multi-master protocol based on serial communication that allows
communication between two microcontrollers. There are also libraries that facilitate
communication and handle all lower layers. Multi-master means that all connected
microcontrollers can send data when needed. This is the main difference from the master-slave
protocol, where only the master device can send data. The Arduino Uno has a UART that can be
used from a USB cable. RX or TX pin can also be use. The Arduino Zero has only one USB port
for serial communication. The Raspberry Pi can connect many microcontrollers to the USB port.
GPIO (RX0 / TX0) can also be used.

45
 Figure 3.16(USB communication)

This connection are made when we are using the USB Port for communication between two
boards.

Figure 3.17(Communication through Rx and Tx)

This connection is made when we are using the general input and output pins for communication
between two micro controllers.

46
 PCB Design

Figure 1.18(PCB Design)

This PCB is designed by using the proteus. The tracks are defines by blue lines. There are t-
blocks, Arduino, Buck converter, Moter Driver IC, MOSFETs.

Resistors
Resistor puts resistance in the system. They hinder the flow of the current in the system. They
are use to regulate the signal, limit the current flow, divide the voltage.

Figure 3.19(Resistors)

47
Resistor of 270ohm and 2.7k ohm are use in hardware part of this project.

LEDs

LEDs are used for the lighting purpose. They are connected on the top with the 12v battery. They
helps to capture the good quality image having less amount of noise in it. Image should be clear
so that it gives us maximum accuracy.

Figure 3.20(LEDs)

48
 Chapter 4

Design (S/W)

4.1 Arduino

Arduino programing language s based on C and C++ abstraction. The programs written on
Arduino IDE are called sketches. After a sketch s compiled on Arduino DE t s then uploaded to
Arduino board.

Figure 4.1(Arduino IDE)

Arduino has the open-source libraries those are available for everyone. Its makes coding of
hardware easier.

Features

 inexpensive.
 open source in hardware.

49
  do not need to external programmer (Burner)
 programming ease.
 open source in software.
 IDE Software operate on any operating system.

4.2 Proteus

Simulation of this project s done on proteus. PCB design of this project s also done n proteus.
Proteus s used to simulate and design electronic circuits. It was developed by the LaCenter
electronics. By using proteus you can make two- and three-dimensional circuit designs as well.

4.3 Schematic

Figure 4.2(Proteus Simulation)

Motor and Solenoid are connected with IR sensors and a controller. When first sensor gives
input, the roller will start to move and when second sensor gives input solenoid loads the ballot
paper in the machine. This process will continue in loop till no ballot paper left inside the
machine.

50
4.4 PCB Design on Proteus

Figure 4.3 (PCB)

This is the PCB layout of hardware design and made on Proteus. This design contains the
following components two 270-ohm resisters, two 2.7k resistors, motor driver IC,
MOSFET(IRLZ44n), Arduino UNO and bug convertor.

4.5 PyCharm

The second half work of this project is done on PyCharm. Code s written on PyCharm using
image processing library Open Cv.
PyCharm s an DE of Python which provides programmers with a wide range of tools, libraries
and s tightly integrated to create a working environment for productive Python, Web, and data
science development.

51
PyCharm Features

 Intelligent Coding Assistance. PyCharm provides intelligent code completion, code

inspection, on-the-fly error highlighting and quick fixes, and automated code refactoring
and rich navigation capabilities.
 Built-in Developer Tools.
 Web Development.
 Scientific Tools.
 Customizable and Cross-platform DE.

PyCharm also allows programmers to split longer classes and methods by extraction methods.
PyCharm makes t easy to create a variety of web applications n Python that support widely used
web technologies such as HTML, CSS, JavaScript, TypeScript, and Coffee Script.

Figure 4.4(Py Charm)

Py Charm is easy to use. It allows user to use the different libraries, those are open source and
available for everyone. Its user interface is also incredibly good that helps in coding.

52
4.6 Open CV-Python

OpenCV s a python and C Library that is used to solve computer vision and image processing
problems. We used python version because t s general purpose programming language and
extremely popular because of its simplicity.

4.7 Flow Chart

53
 Chapter 5

Mechanical Design

5.1 OMR (Optical Mark Recognition)

This project is based on OMR mechanism. A special type of paper is entered from one end of the
machine and exit from another end. During this the paper is scanned and this scan is use for
further analysis.
In our case, we are taking the image of the paper instead of scanning it. After this image is saved
in a storage and use for image processing purpose.
It is widely use where the large numbers of candidate apply to check the mcqs sheet and evaluate
the results immediately.

Figure 5.1(OMR)

Uses
OMR can be use in the following services
 Elections
 Surveys
 Banks
 Examination
 Evaluation

54
Advantages
OMR has the following advantages
 A special type of paper and ink required for its working so there are less chance of
getting the fake papers.
 It takes less time to conclude the results
 There are less chance that people may question about credibility of the machine.

5.2 Limitation

OMR has the following limitations

 There are chances of more than one paper comes out of machine at same time
 Cannot gather large amount of text, it will complicate the results
 Data loss
 Result might be tempered easily by any insider

OMR is widely use as a MCQS checker, but we can also use it for election purpose. This project
uses OMR mechanism for counting votes. Our OMR is simply combination of motor, solenoid,
and a camera.
First of all, the ballot paper are loaded inside the machine, when the camera takes the image it
sent the signal to solenoid to load the paper in the roller. Once the paper is loaded in the roller
the motor starts to rotate and paper comes out from second end of the machine. After this camera
takes the second image and this process will continuous until there is no ballot paper is left inside
the machine.

5.3 Design
The design of the machine is made by using Acrylic Sheets. The width of sheets is 10mm which
give the machine enough hardness so it will not get break. Five iron rods are used to provide the
support to the sheets to stick them together. The sheets are transparent so that we can see the
working of machine from all sides so that we can easily find the errors and conclude them. The
height of acrylic sheets is 24cm and length is 42cm.

55
 Figure 5.2(Acrylic Sheets)

There are two IR sensor connected to the bottom of the acrylic sheet, they are use for input from
the ballot paper. If there is a ballot paper inside the machine, they give the signal to the Arduino.

Figure 5.3(Mechanical Design)

56
The top is also made of acrylic sheet. There are LEDs connected on the top and the camera.
LEDs helps us to get clear and noise less image. There is PCB inside to this which is attached to
acrylic sheet in the middle. PCB contains all the hardware require to run this project.

Figure 5.4(Mechanical Design)

There is a motor and solenoid connected with the acrylic sheet and with the metal rod. Motor and
solenoid are use for ballot paper input and output mechanism. Motor is connected with the roller
by the help of a rubber that’s help roller to rotate. When the motor rotates the rubber connected
with roller and motor also rotates. At the bottom there is the spacer that’s helps ballot paper to
come out once at a time, its is not enough efficient there are limitation. When the solenoid
switches it makes pressure on the spacer and the next ballot paper is loaded inside the spacer.
The distance between the spacer is 0.7mm, that helps the spacer to not load more than one ballot
paper in same time.

Figure 5.5(Iron Rods)

There are five metal rods, use to tightly stick both sides of acrylic sheets together. The roller is
also connected to the metal rod so that it provides enough force to the ballot paper and the ballot
paper comes out. There are bolts on both side of metal rod. They are tightly closed, if they are

57
loosely closed the space between the space can also be changed and have the effect on ballot
paper in or out mechanism.

Figure 5.6(3D printed Cap)

It is 3D printed cap inside which the rod with the roller is connected. This helps roller to rotate
smoothly.

Figure 5.7(3D printed Cap)

It is also 3D printed, used to provide the support to the paper tray where all the ballot papers are
loaded. It allows paper tray to move freely.

58
6.1 Expected Outcomes of the Project/Benefits:

The expected outcomes of our project include:

 Developing a Vote counting machine that so that a lot of workforces s not wasted on the
repetitive task.
 Eliminating human error from vote counting process.
 Making the vote counting process more secure.
 Making the voting process more efficient without changing the entire system.
 Making the vote counting process faster.
 It can also be used to count VVPAT slips.

6.2 Conclusion

Voting technology promises faster elections, more accurate voting, and more accurate ballot
counting. Efficacy, efficiency, and satisfaction are three suability metrics used to assess the
usability of any electronic voting system. Security and non-repudiation are important aspects of
the voting framework to ensure classification and validation of the voting framework. Error
checking should also be considered when evaluating the vote exclusion system. Speed and
accuracy can be guaranteed by using the e-casting ballot system.

59
References

[1]https://www.pdfdrive.com/practical-machine-learning-and-image-processing-for-facial-
recognition-object-detection-and-pattern-recognition-using-python-d188718832.html

[2]https://cdn.sparkfun.com/datasheets/Dev/RaspberryPi/RPiCamMod2.pdf

[3]https://pdf1.alldatasheet.com/datasheet-pdf/view/17807/PHILIPS/IRFZ44N.html

[4]https://www.tutorialspoint.com/dip/image_processing_introduction.htm

[5]https://www.analyticsvidhya.com/blog/2021/09/some-amazing-applications-of-opencv-
library/

[6]https://www.sunrom.com/p/infrared-sensor-board

[7] Omachi, S., & Omachi, M. (2007). Fast template matching with polynomials. IEEE
Transactions on Image Processing, 16(8), 2139-2149.

[8] Hisham, M. B., Yaakob, S. N., Raof, R. A. A., Nazren, A. A., & Wafi, N. M. (2015,
December). Template matching using sum of squared difference and normalized cross
correlation. In 2015 IEEE Student Conference on Research and Development (SCOReD) (pp.
100-104). IEEE.

[9] Swaroop, P., & Sharma, N. (2016). An overview of various template matching
methodologies in image processing. International Journal of Computer Applications, 153(10), 8-
14.

[10] Smith, E. H. B., Goyal, S., Scott, R., & Lopresti, D. (2011, September). Evaluation of
voting with form dropout techniques for ballot vote counting. In 2011 International Conference
on Document Analysis and Recognition (pp. 473-477). IEEE.

60
[11]https://en.wikipedia.org/wiki/Elections_in_Pakistan#:~:text=At%20the%20national%20level
%2C%20the,chosen%20by%20elected%20provincial%20legislators.
[12]https://pakvoter.org/the-electoral-system-in-pakistan/
[13]https://pide.org.pk/research/electoral-voting-machines/
[14]https://www.diva-portal.org/smash/get/diva2:1018802/FULLTEXT01.pdf
[15]https://cscr.pk/explore/themes/politics-governance/electoral-system-of-pakistan-and-its-
evolution-proportional-representation/
[16] https://www.researchgate.net/publication/261394443_Electronic_voting_machine_-_A_review
[17] https://pide.org.pk/rasta/wp-content/uploads/Hina-Bint-Haq-Conference-Paper.pdf
[18] https://ieeexplore.ieee.org/document/6208285
[19]https://www.researchgate.net/publication/320927313_Raspberry_PI_and_Arduino_UNO_Working_T
ogether_as_a_Basic_Meteorological_Station

61