Volume 5, Issue 5, May – 2020 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165

Number Plate Recognition a for

Parking Assist and Road Safety
Senthil Pitchappan V Sree Harish Rajan A R
BE in Electronics and Communication Engineering BE in Electronics and Communication Engineering
Velammal College of Engineering and Technology, Madurai Velammal College of Engineering and Technology, Madurai

Abstract:- As an Engineer and an Indian citizen, police officers we use machine learning to lodge
Parking is a major problem in our country as well in complaints on the infatuate.
our surroundings such as Apartments, Educational
Institutions, etc. and also road Safety. This problem can Parking is an imperative element within any
be autonomously monitored. Automobile parking is a transport structure, whereby automobiles obligated to be
predominant factor within any transit system, whereby parked at every terminus. Most automobiles are parked
automobile are habitually parked at terminuses. With maximum of the time throughout the day. Parking
an escalated number of motor vehicles on roads expediency is the luxury of accessing a benign and
peculiarly in emergent countries, there is a need for an desirable parking location, hence distressing the ease of
automobile identification mechanism that is operative, reaching destinations. India which sees itself as a hopeful
affordable, and competent. In most academic global influence and an economic pivot can accomplish its
institutions, apartments, in addition car parks, the goal if and only if a huge amount of women folk
constant car park entry registration process for visitors, participates in the advancement progression. This paper
faculty or students entering the institution involves a presents an analysis review on the principal need of the
security guard having to authenticate membership intelligence security systems with technology requirements
particulars by examining for membership sticker on the and challenges to build the system. Since the prediction of
windscreen of the vehicle or by checking the driver's such an incident is not possible hence to curtail the prospect
credentials and identification card. This procedure of of physical violence (robbery, sexual assault etc.) by
inscription is dreary and time-intense and is disposed to possession of all the help tools ready to securely escape
to erroneous recordings, also, the backup and from insurgent situations. This condenses the hazard factor
distribution of this vehicle information are grim because and brings assistance when needed. The social networking
the information is rigid to duplicate. This procedure can is the part of our life and also a source for women
be automated using Image Processing and a dedicated harassment by uploading the offensive photograph taken by
Android Application for owners and Policemen with hidden cameras, even though these cases might happen
automatic theft control, Email, and messaging. Road with innocence males, in some such cases these guys end
Safety on the supplementary side is also completed their life by committing suicide.
using the identical application where the user can fail
grievances concerning traffic violations, accidents, Automobile car parks are an imperative element
women harassments etc., which will be automatically within any conveyance system, whereby automobiles are
updated in the police application. And he/she can take often parked at destinations. With an increased number of
necessary decisions. Concluding, this proposed project motor vehicles on roads especially in emerging countries,
can bring a new revolution in parking assist as well as there is a need for a automobile credentials verification
road and women safety. mechanism that is operative, inexpensive, and competent.
In most academic institutions, apartments, and automobile
Keywords:- Image Processing, Spyder, Firebase, MySQL, parks, the constant car park entry record-keeping process
MIT App Inventor, Android Application, Raspberry Pi, Pi for guests, faculty or students entering the institution
Cam/Webcam, Tableau. involves a security guard having to confirm membership
credentials by checking for affiliation marker on the
I. INTRODUCTION windscreen of the vehicle or by checking the driver's
credentials card. This process of writing is dreary and stint-
Vehicles are very dense in a populous country such consuming and is prone to erroneous recordings,
as ours. The idea is to create a smart solution to deal with furthermore, the backup and distribution of this vehicle
modern problems. As there are many vehicles the information are grim because the figures are hard to
government needs to monitor these vehicles for various duplicate.
parameters. It is nearly impossible to man the streets to
catch up with the violators. So here we bring up our
research on Image Processing of Number Plates, here using
the image processing technology we are able to scan the
number plates of the vehicles and identify the violators and
file infractions on the person. So instead of using separate

IJISRT20MAY872 www.ijisrt.com 1722

Volume 5, Issue 5, May – 2020 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
II. PROGRAMMED NUMBER PLATE the image acquisition stage. Image pre-processing mainly
DETECTION involves translating the RGB image into grey color, noise
removal, and border enhancement for brightness. Image
Programmed Number Plate Recognition is a method pre-processing is usually done through image filtering.
where the automobiles are branded or recognized using
their number plate or license plate. PNPR uses image image = cv2.imread('C:/Users/N/Desktop/Test.jpg')
processing procedures so as to the vehicle number
plate from alphanumeric images. PNPR systems normally B. Grey Scale Conversion
comprise of two components: A 5mp camera with infrared In the 1st step we obtain the saved image from the
and night vision is used in capturing of vehicle number camera image or an image from local files. Greyscale
plate images, and software that extracts the number plates image is preferred because it displays the required part
from the captured images by using a character recognition clearly without deflecting any minor unnecessary itinerary.
tool that allows for pixels to be translated into numerical Grey scaling an image is important because it surely does
readable characters. A license plate recognition system make the processing system easy to read the image and
generally works in four main parts namely image follow further instructions easily without any difficult
acquisition, license plate detection, character segmentation, hurdles that trouble the processor in processing. Next we
and lastly character recognition. It is a sequential process convert the image into a grayscale image using Open-CV, a
where first the original RGB image gets captured then the python-based image processing library file.
image gets converted into grayscale that is, it gets cv2.imshow('Gray image', gray)
binarized. Then the image is filtered of noise and such
similar undesired components. After the image is filtered C. Filtering of Image(blur)
the license plate image is specifically cropped out from the A bilateral filter is used for smoothening images and
main image. After magnifying and selecting the image it reducing noise, while preserving edges. It is easy to note
then gets read by character segmentation that is done using that all these denoising filters smudge the edges, while
open cv software where the image is considered and a Bilateral Filtering retains them. Image blurring is achieved
separate layout is taken where the image gets segmented by convolving the image with a low-pass filter kernel. It is
structure by structure until the whole segment is due useful for removing noise. These noises tend to infect the
processed. image with interference which may lead to erroneous
readings. It removes high-frequency content (e.g: noise,
edges) from the image resulting in edges being blurred
when this is filter is applied.

Blur = cv2.bilateral(gray,11,90,90)

D. Canny Edge Detection

Canny Edge Detection is a popular edge detection
algorithm. It was developed by John F. Canny in 1986. It is
a multi-stage algorithm. OpenCV puts all the above in a
single function, cv2.Canny(). We will see how to use it.
Fig 1:- shows a typical PNPR process. The first argument is our input image. The second and third
arguments are our minimal and maximal values
A. Image Acquisition respectively.
The first step is the image acquisition phase. The
image of the vehicle is apprehended using a photographic Edges = cv2.canny(blur,30,200)
camera. The constraint is that the image of the vehicle
should be taken in such a mode that the selected input E. Contouring
image comprehends the rear or front view of the vehicle Contours can be explained simply as a curve joining
with the number plate. This procedure can also be done by all the constant points (along the boundary), having the
capturing a video and absconding the image frame by frame matching color or intensity. The contours are a beneficial
to obtain the image. So by fixing suitable cameras at proper tool for shape breakdown and object detection and
terminuses we are definitely able to apprehend any said recognition. For better accuracy, practice binary images.
image through video processing. The image is usually Contour basically just crops out the essential part from the
captured in an RGB (Red, Green, and Blue) color model. unabridged image. It has turned out to be very valuable
The captured image is affected by many factors like because it unbiased traces along the boundaries and legibly
photosensitive system, distortion, system noise, deficiency provides very accurate outcomes. So before finding
of exposure, or extreme relative motion of camera or the contours, apply threshold or canny edge detection. find
automobile thus resulting in a dilapidation of a captured Contours function modifies the source image. So, if you
vehicle image hence unfavorably affecting the outcomes of want source image even after finding contours, already
the inclusive image processing. As a modification store it to some supplementary variables. In OpenCV,
mechanism, an image pre-processing stage is familiarized finding contours is like finding white objects from a black
to take care of any blunders that may have occurred during

IJISRT20MAY872 www.ijisrt.com 1723

Volume 5, Issue 5, May – 2020 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
background. So, remember, the object to be found should The above code uses pytesseract package to convert
be white and the background should be black. image file to string file automatically.
cnts, new = cv2.findContours(edges.copy(), cv2.RETR_
LIST, cv2.CHAIN_APPROX_SIMPLE) III. PARKING ASSIST

F. Number plate cropping Each and every number plate is processed using image
It is also called arc length. It can be found out processing in Spyder (Python) or MATLAB. Every car has
using cv2.arcLength() function. Second argument specify its own parking lot in an area. Our project helps the drivers,
whether shape is a closed contour (if passed True), or just a as well as the security, correctly park in the respective
curve. place. While the car enters the parking lot, the car number
plate = None is obtained. This car number is correlated and compared in
for c in cnts: the database and finds the correct parking lot number. The
perimeter = cv2.arcLength(c, True) parking lot number is delivered to the driver and he can
edges_count = cv2.approxPolyDP(c, 0.02 * perimeter, T park in a specific lot. Even though, if the driver parks in the
rue) wrong lot, it intimates the driver as well the security by a
if len(edges_count) == 4: notification. Car in time and out time, Car number is
x,y,w,h = cv2.boundingRect(c) uploaded to the database and is also intimated i.e. notified
plate = image[y:y+h, x:x+w] to the owner through a web app and android application.
break Automatic Bill is generated using in time and out the time
cv2.imwrite("plate.png", plate) of the car.

The above code crops only the number plate in the Parking lots in a specific venue can be pre-booked
image processed as saves the copped image as plate.png using the Web Application. If there is any malpractice done
by the driver, police accounts are also integrated with the
G. Character Segmentation web app, so that theft and other cases can be filled and FIR
Python-tesseract is an optical character recognition (if the case is serious) can also be obtained as a PDF in the
(OCR) instrument for python. That is, it will distinguish Web Application itself. The output from Python is updated
and “read” the text entrenched in images. Python-tesseract to My SQL database using Python. PHP is used to retrieve
is a wrapper for Google’s Tesseract-OCR Engine. It is also data from SQL Database and view it in a Web Application.
useful as a stand-alone supplication script to tesseract, as it This Web Application will be able to send E-Mail,
can read the entirety of image types supported by the Message, Notification to the owner as well as a nearby
Pillow and Leptonica imaging libraries, including jpeg, police station in a case of emergency. The owner can
png, gif, BMP, tiff, and others. Additionally, if used as a him/her self-update the GPS location, lot number, and in
script, Python-tesseract will print the recognized text as an time & out time in her App, pre-book parking lots as at
alternative of writing it to a file. times people won’t be using Drivers. A web application
import pytesseract will be provided to the drivers, to update the parking lot
text = pytesseract.image_to_string(plate, lang="eng") number, in & out time, GPS location, and other information
print(text) are to the owner. Furthermore, Tableau can be used to
visualize the data sets in the server directly in the software
using graphs, time logs, time periods, payment, etc.

Fig 2:- Women Safety System

IJISRT20MAY872 www.ijisrt.com 1724

Volume 5, Issue 5, May – 2020 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
IV. WOMEN AND ROAD SAFETY

This Web Application is not only used for Parking

Assist, It can also be used in the case of Road Accident,
Traffic violations, etc. The user can enter the Car Number
with the GPS location and the fault they did including the
locality. Then the police officials can do the necessary.
When a woman is being harassed or being followed by a
stranger who is doubtable by the women, she can send the
GPS information, SMS to nearby police stations and
relatives by a press of a button. This also helps women in a
needy and most followed issue, the application is able to Fig 5:- Edge image
display the details and location of all major vehicle repair
shop in a locality with best customer reviews and women-
friendly ones in case of vehicle breakdown and it is also
coded to notify the selected contacts when in the case of
vehicle breakdown so that they can know their current
location and also track the progress of them repairing the
malfunction. This ensures the maximum safety of women
when struck in any kind of remote areas where their safety
is at peril. It basically works like a security application for
women can select emergency contacts from their contact
list and save them who when they’re in need of assistance
they can simply just at the press of a button call for help.
Fig 6:- contour image
An application is developed to lodge complaints about
missing vehicles and we also provide a follow up with
police and the user is able to see the status of the lodged .
complaint through this app. With the help of this Fig 7:- Final cropped image after character segmentation
application the person who lodges the complaint need not
have the use to often go to the police station and check up
on his complaint, they can simply use the application.

V. RESULTS

Fig 8:- Final output in Spyder IDE

VI. CONCLUSION
Fig 3:- grey scaled image
Problems and challenges associated with car park
vehicle identification details formed the basis for this
research. The main aim of this research was to develop an
automatic number plate recognition system for car park
management, using Optic Character Reader (OCR) on a
mobile device. The OCR process forms the basis of the
entire system that was proposed by the researcher as a
means of tackling the challenges faced by the security
guards during the vehicle entry registration process. The
completion of the proposed system resulted in the
following benefits:
 Elimination of the hard copy occurrence book and the
Fig 4:- Blurred image need to have to physically write onto the book, because
all the vehicle details records will be digitized.

IJISRT20MAY872 www.ijisrt.com 1725

Volume 5, Issue 5, May – 2020 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
 Hastening of the car park vehicle identification process [10]. Automatic number plate recognition for the
including the entry and exit process, thus shortening the observance of travel behavior. Retrieved from
time duration. http://www.isv.uni-stuttgart.de/vuv/publication
 Accurate recording of vehicle information. /downloads/200805_Fr_PJ_JS-ANPR.pdf
 Provides a means of easy information sharing and [11]. Lotufo, R., Morgan, D. & Johnson, S. 2013.
International Journal of Computer Applications (0975 – Automatic
8887) Volume 175 – No.7, October 2017 42 [12]. license plate recognition (ALPR) a state-of-the-art
information backup. review. Journal of IEEE transaction on circuits and
 Real time information sharing of the vehicles entering system for video technology, vol. 23, no, 2013, pp.
and exiting the Institution to the Head of Security. 311-325 DOI:10.1109/TCSVT.2012.2203741
 Easier analysis of the vehicle information captured. [13]. Camera-sdk Retrieved from http://www.camera-
[14]. sdk.com/p_89-how-to-implement-number-plate-
The process what generally occurs here is that first the [15]. recognition-in-c-onvif.html
image gets captured. Then nominally the image is [16]. Reshma, P. 2012. Noise Removal and Blob
converted into grayscale. After the grayscale conversion we Identification Approach for Number Plate
need to filter the image of blurs and another synchronicity Recognition. Retrieved from
error in the image by processing it. Then the image gets http://research.ijcaonline.org/volume47/
cropped only selecting the number plate, this process is number8/pxc3879992.pdf
specifically done to obtain precise and correct results. The
process for cropping the image must be very carefully
coded because the software shouldn’t erroneously select
any other number present in the image. After the image is
cropped and correctly obtained, we need to detect the image
and send it for processing further operations. The cropped
image simply doesn’t cope, first the image gets segmented
through character segmentation so the optical character gets
segmented. We basically make the machine learn to detect
what alpha numeric character is directed and displayed in
the processed image. By doing so we can automate the
entire process so even a simpleton like a layman just like
that is able to perform these tasks by the press of a switch
or button.

REFERENCES

[1]. Litman, T.2013. Parking Management Strategies,

Evaluation and Planning. Retrieved from
http://www.vtpi.org/park_man.pdf
[2]. Subraman T. 2012. Parking Study on Main Corridors
in Major Urban Centre. International Journal of
Modern Engineering Research (IJMER) ISSN:
2249-6645. Retrieved from
[3]. http://www.ijmer.com/papers/vol2_issue3/AE237427
48.pdf
[4]. Cornwall. 2009. Drivers on police files for life.
Retrieved from http://www.westbriton.co.uk/Drivers-
police-files-
[5]. life/story-11398048-detail/story.html
[6]. Roberts & Casanova 2012. Automated License
Plate Recognition Systems: Policy and Operational
Guidance for Law Enforcement. Retrieved from
https://www.aclu.org/files/FilesPDFs/ALPR/federal/N
HTSA/15948-16075DOJ-IACP%20report.pdf
[7]. ACPO (2013). The police use of Automatic
Number Plate Recognition Retrieved from
[8]. http://www.acpo.police.uk/documents/crime/2013/201
303CBA-ANPR.pdf
[9]. Friedrich, M., Jehlicka, P. & Schlaich, J. 2008.

IJISRT20MAY872 www.ijisrt.com 1726