SHAHEED BHAGAT SINGH STATE UNIVERSITY

FEROZEPUR

PROJECT TITLE

VEHICLE DETECTION AND COUNTER SYSTEM

SUBMITTED BY –

ABHINAV SINGH(2007753)

AKASH KUMAR(2007755)

ANMOL SINGH(2007757)

(DATA SCIENCE )

( 3rd YEAR)

Batch 2020-2024
 Table of Contents
Chapter 1. Introduction
1) Introduction to the System
2) Abstract
3) Aim
4) Objective
5) Goal
6) Limitation

Chapter 2. Hardware and Software requirement

1) Introduction
2) System environment
3) Software requirement
4) Hardware requirements

Chapter 3. System Analysis

1) Purpose
2) Project Scope
3) Existing System
4) Proposed System
5) System Description

Chapter 4.Implementation
1) Python
2) OpenCv(Cv2)
3) Glob
4) Numoy

Chapter 5. Output

Chapter 6. Conclusion
 CHAPTER-1

Introduction:

In this chapter we mentioned the software and hardware requirements, which are
necessary for successfully running this system. The major element in building systems
is selecting compatible hardware and software. The system analyst has to determine
what software package is best for the “Vehicle Detection System” and, where
software is not an issue, the kind of hardware and peripherals needed for the final
conversion.

In this chapter we introduce the background of the thesis, and the focus area. Furthermore
the problem descriptions and research questions, research aim and limitations.

Vehicle detection is a computer vision technique that makes a system able to locate objects
in an image or video. This bounded box[23] from the segmentation can help a human or
driver to identify and locate the objects in a matter of seconds and quicker than
unprocessed images. Therefore, the goal of object detection is to implement this
intelligence to a computer.
There are different techniques to perform object detection. Popular deep learning–based
approaches using convolutional neural networks CNNs, such as YOLO and SSD, which
will automatically learn to detect objects within frames. To explain further on object
detection, more knowledge on machine learning is presented below. The Figure 1 below
shows the difference between image recognition and object detection.

Abstract
Object detection exist in many countries around the world after recent growing interest
for autonomous vehicles in the last decade. This paper focuses on a vision based approach
focusing on vehicles and pedestrians detection in real-time as a perception for autonomous
vehicles, using convolutional neural network for object detection. A developed YOLOv3-
tiny model is trained with the INRIA dataset to detect vehicles and pedestrians, and the
model also measures the distance to the detected objects. The machine learning process
is leveraged to describe each step of the training process, it also combats overfitting and
increases the speed and accuracy. The authors were able to increase the mean average
precision; a way to measure accuracy for object detectors; 31.3% to 62.14% based on the
result of the training that was done. Whilst maintaining a speed of 18 frames per second.

Research aim
The research aim of this thesis is to detect vehicles and pedestrians in a real-time environ-
ment with fast image processing of 20-25 frames per second (FPS)

Problem descriptions and Research Questions

An object detection algorithm will be built with many neural network layers with different
purposes. A developer choose an algorithm and develop it according to the usage purpose.

Limitations
Some limitations in this type of area can occur if a low resolution camera is used which
may result in bad image quality or if a computer with low (GPU) graphics processing unit
is used which may lead to slower object detection process. To train a new dataset requires
a effective GPU, in our case we are using AMD RX 580 graphic card. In order to speed
up the process a better performing GPU is needed otherwise, the time required to train
our dataset will be longer than usual. Also to achieve better FPS (Frame per second), the
GPU plays a fundamental role.
When two people stand close to each other it can become difficult for the algorithm
to detect it correctly. Most commonly the algorithm generate a bigger bounding box and
interpret it as one person. However, there is still a notification of pedestrian detection to
the driver by the developed system in such circumstances.
 CHAPTER-2

System Environment:

After analysis, some resources are required to convert the abstract system into the
real one.

The hardware and software selection begins with requirement analysis, followed by a
request for proposal and vendor evaluation.

Software and real system are identified. According to the provided functional
specification all the technologies and its capacities are identified. Basic functions and
procedures and methodologies are prepared to implement. Some of the Basic
requirements such as hardware and software are described as follows: -

Hardware and Software Specification

Software Requirements:

• Technology: Python
• IDE :VsCode
• Client Side Technologies: command line interface
• Server side : python technologies
• Operating System: Microsoft Windows

Hardware Requirements:

• Processor: Pentium-III (or) Higher

• Ram: 64MB (or) Higher
• Hard disk: 80GB (or) Higher
 CHAPTER-3

System Analysis

Project Scope:

In the modern age. Many people have vehicles. Vehicle is now a basic need. Every place is
under the process of urbanization. There are many corporate offices and shopping centers
etc. There are many recreational places where people used to go for refreshment. So, all
these places need a parking space where people can park their vehicles safely and easily.
Every parking area needs a system that records the detail of vehicles to give the facility.
With the help of this system we can deliver a good service to customer who wants to park
their vehicle into the any organization’s premises.

The project has a wide scope, as it is not intended to a particular organization. This
project is going to develop generic software, which can be applied by any businesses
organization. More over it provides facility to its users. Also the software is going to
provide a huge amount of summary data.

Proposed System:

Vehicle Detection system is a web-based technology that will manage the records of
the incoming and outgoing vehicles in a parking house.

System Overview:

The key features required in the system are as follows:

Dashboard: In these sections, admin can briefly view the number of vehicle entries in a
particular period.

Category: In this section, admin can manage category (add/update).

Add Vehicle: In this section, admin add vehicle which is going tofast.

Manage Vehicle: In this section, admin can manage incoming and outgoing vehicle .

Reports: In this section admin can generate vehicle entries reports between two dates.
 CHAPTER-4

Implementation issues

PYTHON

Python is a widely used general-purpose, high level programming language. It was initially
designed by Guido van Rossum in 1991 and developed by Python Software Foundation. It
was mainly developed for emphasis on code readability, and its syntax allows programmers
to express concepts in fewer lines of code.Python is a programming language that lets you
work quickly and integrate systems more efficiently.

CV2

Opencv is an open source library which is very useful for computer vision applications such
as video analysis, CCTV footage analysis and image analysis.

cv2 is the module import name for opencv-python, "Unofficial pre-built CPU-only OpenCV
packages for Python". The traditional OpenCV has many complicated steps involving
building the module from scratch, which is unnecessary. I would recommend remaining with
the opencv-python library.

GLOB

The glob module is a useful part of the Python standard library. glob (short for global) is
used to return all file paths that match a specific pattern.

We can use glob to search for a specific file pattern, or perhaps more usefully, search for
files where the filename matches a certain pattern by using wildcard characters.

According to Wikipedia, “glob patterns specify sets of filenames with wildcard characters”.

These patterns are similar to regular expressions but much simpler.

• Asterisk (*): Matches zero or more characters

• Question Mark (?) Matches exactly one character
NUMPY

NumPy is a Python library used for working with arrays.

It also has functions for working in domain of linear algebra, fourier transform, and
matrices.

NumPy was created in 2005 by Travis Oliphant. It is an open source project and you can use
it freely.

NumPy stands for Numerical Python.

In Python we have lists that serve the purpose of arrays, but they are slow to process.

NumPy aims to provide an array object that is up to 50x faster than traditional Python lists.

The array object in NumPy is called ndarray, it provides a lot of supporting functions that
make working with ndarray very easy.
 S1ojeflamning& tstimahen C ponamly
Poject win co xtmghibm s nolernte=pon
intdiaddy olen ka tbulity 3-uoly
ad beer h
nfg
SHortvg
h uirCm&nt onalysis and rej ipbmwndotun
ph
ojeckanwing tvelus eimecion
Sauuro
chorat Heuties and
pa1ew o
o Frject
Cwd thow plowuwgt j
octiuiios bond
Hceyet extimates. Th
thot Mud to
ou bkevwt po-wdors
pohuwdons
be es-fimostol
intduolu
PSz EstimahoM
Cest Esthimaan
EFRort Estimotion
Dunuethon ITiwe Eshmin
(1) Pject wnolkorto Sing
Siz
s k wpevtend poamodn o Ony
projec cmo sou e ehmodrel
be-re
a.ctu implivmont otievn
Siu prejeet Con b
mujwd i n Lim Coel (LOc) by leldy
nUmben ws Co olu obnulo
th pojeet . ws Coebu dapend vpavw Geding
Pe ractites amd may vy OCCY di to

unetien otHy t h rject. 3+ is

diffiadse
odomint Sin CoCoo beort Cedlig pho
RXact
wh
& LOCCm reciefeol by Componi
Siilon. ezistiy oyects s Fex tyomp

RCeob e t T olmlupud by Kontamo Fuhvchi

in c Jnyyng koe abou 6 Foo Li Ceole
in h Jibrey. his odu as Confusi with
OVen So þubkiu Afl Pone iovMs.
In

phythom Onol
duiSol in iooni ts hunu h
maximum is Cool opoximodey ktwun
fooSo u Cool. This IS

Omel Small
bojt Hime the Si
Siz
t pjeet s Small anol
s eCeo isJu viatl.
e)EHo Estima
Efft eimatien is
Hkeimsien {fris
n Atws peumo and mon - ho

1fo»ts quirtd to borcat th Softusot pacjett,

For rt estimatien Softuou si shsulel

be nowww f is obey ha ho Cdeustu.

This Can also Calrwattol by histoi dota.

Amevn dabow tot wiL be uiesl t
Cewnplate a tosh. S4
COco M0 Mo clul Cax be Glevrd

by ermola E a CkLoc)b PM
b10S
KLOC
Ne.No. s ecda (Ku)
moyI mum Muwen
iny
Co = So0 LoC 5 KLoc
Twn
Efto4s = 240.5) PM

Eforty 1IS 9 fersu- Month

his pmjeet O Ce urter tru is
wallis at o 9+ is a ml bojeet.
Hmet
Poee nuwbor
opoys coulupuond shoul
nite.. No oy coliio btuet t v o
claulbpon
3 Estimaion
Onee si on 6ts esticdssl
Cost the Sofee
Ctan be
estiweded T pnjectbuelr
Cia ivdivect CLno ollrect Cests o
esfimedts t ua Cos daliui
CAnd
ut 2ualrty
ex pecfathons. "\at iany, whetn
you!
oingt to mopth projeet
Suctful. w thveugh th ct e thimsdion
Pascen nCost exhmodion Vnious facf
incu Coy, Cos for oduepey,
Mantww C-s Cos 6 uiro toals.
Th cLe 3 Hechn' zuy
Cost estimntian
sg Enpirical estimoction fcehuijue
which h Covt wi portoied aol besee by
Erpu bostel
2ws on albta thot e s b - e t is
Cellee fes ueuieuy nem proyet cend also
boseol Som
quyss is techni gu o
tsiz . j tef th
Cost. D to hamolw su and Stw o yvjr ds
Cano also obulapovs o
yunl by
 4.leost 50 Cst's
elulo

Tim Estdto n Coun Calelt th

ov ryeet
Hen Coulculspion f si ounol efor sts.
Th Sum tHwe uvired to Cemhte oll

tospstn th torfol

. Sf rsolu
inwsfes Ho Cewplate projecf

Cales ouwtien rjecf. COun pajtut s

is

siu nu sfiatëm ouwuflm

hvinSmall
be lrs. Estimote
i m wi
th
Gn
Con b Culotol by
T a (E3Ho)
Smll
ponjecds
b 0 38
38
T 2-S IISs
26 Months
CHAPTER 5: OUTPUT
 CHAPTER-6

Future Scope & Conclusion

FUTURE SCOPE

Thisweb application involves almost all the basic features of the online vehicle parking system. The
future implementation will be online help for the users and chatting with website administrator.

CONCLUSION

The project entitled “ VehicleDetection System” is developed using Python, Cv2,Blob and as front
end and Python and vscodein back end to computerize the process of company visitor. This project
covers only the basic features required.

After research and compression, results showed that the You Only Look Once (YOLO)
algorithm is the most suitable for solving real-time object detection in traffic environment.
In this thesis a version of the YOLO entitled YOLOv3-tiny is developed and presented
with higher mean average precision on vehicles and pedestrians detection in real-time. The
developed YOLOv3-tiny object detection model is trained with distance measurement to
the detected objects, with the purpose to increase road safety