Smart Parking Using Image Processing To Detect Available Parking Spots

Uploaded by

anshupriya380

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

8 views

Smart Parking Using Image Processing To Detect Available Parking Spots

Uploaded by

anshupriya380

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 4

Proceedings of the Second International Conference on Edge Computing and Applications (ICECAA 2023)

IEEE Xplore Part Number: CFP23BV8-ART; ISBN: 979-8-3503-4757-9

Smart Parking: Using Image Processing to

Detect Available Parking Spots
Alapati Renuka Annam Jitin Chand T. Joy Jeevan G. Geetha
Student, Dept. of Student, Dept. of Student, Dept. of Assistant Professor,
Information Information Information Dept. of Information
Technology Technology Technology Technology
Velagapudi Velagapudi Velagapudi Velagapudi
Ramakrishna Ramakrishna Ramakrishna Ramakrishna
2023 2nd International Conference on Edge Computing and Applications (ICECAA) | 979-8-3503-4757-9/23/$31.00 ©2023 IEEE | DOI: 10.1109/ICECAA58104.2023.10212314

Siddhartha Siddhartha Siddhartha Siddhartha

Engineering College Engineering College Engineering College Engineering College
Vijayawada, India Vijayawada, India Vijayawada, India Vijayawada, India
alapatirenuka28@gm jitinchandannam@g joyjeevan193@gmail geetha.g@vrsiddhart
ail.com mail.com .com ha.ac.in

Abstract: Ineffective parking has become a major areas where the number of vehicles exceeds the
problem for drivers due to urbanization. Particularly number of available parking slots
in urban areas where illegal roadside parking occurs.
For both drivers and administrators, this smart
The first thing a vehicle does when entering a
parking technology streamlines the entire process of parking lot is look for a sign that indicates whether
parking cars. Using a camera to capture images in the lot is completely full, somewhat full, or empty.
real time, smart parking systems employ image Additionally, the driver has no idea where to go to
processing techniques. The system is able to take in order to park his/her car or how many parking
pictures of the parking spaces that convey information slots are present. This results in traffic bottlenecks
about available spaces for parked cars. Images of full around the parking lot entrance and the inefficient
and empty parking spaces in car parks can be use of parking divisions. Ensuring that drivers have
captured by cameras. Instead of wasting time and fuel
access to parking slot information prior to entering
looking for a vacant spot, it is possible to know the
pointing parking lot's empty space by having its
becomes an essential consideration. Using image
photograph. The management of parking spaces can processing, it is possible to count available parking
be crucial in reducing traffic. In our nation, a manual spaces and provide a low-cost solution to the driver's
parking system is used to maintain the majority of the issue.
parking spaces. Ourgoal in this article is to device and
construct a smart car parking system to address the The implementation of this plan will help people
confusion and detect parking spaces using cameras in deal with the issue of increasing traffic congestion,
public buildings, such as offices and shopping centers. time and money, as well as help deliver good open
services, lower vehicle emissions and pollution,
Keywords: Image Processing, parking space
enhance city traffic, increase parking use, and also
detection, camera, OpenCV
prevent wasteful investments. The technology
employs image processing to find the automobile
and also delivers data like the amount of parking
I. INTRODUCTION spaces that are open. The technology uses CCTV
cameras to take pictures, which it then analyses to
Car parking management in most existing car parks
determine how many parking spaces are available.
is currently inefficient and lacks a systematic
In essence, the system is used to plan, analyse,
approach. These car parks are mostly managed create, develop, and test. In each stage of the
manually, which leads to significant time wastage
for users who have to spend considerable amounts
of time searching for empty parking spaces. This
often results in drivers continuously circling the car
park in search of an empty spot, particularly in urban

979-8-3503-4757-9/23/$31.00 ©2023 IEEE 1062

Authorized licensed use limited to: PES University Bengaluru. Downloaded on March 12,2024 at 12:51:52 UTC from IEEE Xplore. Restrictions apply.
Proceedings of the Second International Conference on Edge Computing and Applications (ICECAA 2023)
IEEE Xplore Part Number: CFP23BV8-ART; ISBN: 979-8-3503-4757-9

methodology, image processing techniques will be data to the edge and cloud servers using a software
utilized in the construction of this system. This model made up of sensor, cognitive, network and
system provides information about the number of application layers, the system updated the parking
parking spaces that are available. When drivers lot status and decreased the cost and duration of
enter the parking lot, it will be beneficial to them parking.
all. Due to the minimal number of cameras present,
the system uses image processing. Apart from that, Reference [7] provides a technique for identifying
the system is small and reasonably priced. A occupied and vacant spaces within a parking lot
surveillance camera placed in the parking lot at a while the parking lot is being shot in real time and
suitable height captures an image of the lot. displaying them on an LCD. To evaluate the
condition of each parking place, the study's image
frames were each made grayscale. The converted
image has been cleaned up around the parking area.
II. RELATED WORK Binarization was used to determine the parking
Several models have been proposed by researchers, space's state after that, and the LCD was output with
encompassing diverse frameworks, which can be information about occupied and vacant spots.
utilized for the detection of available parking
spaces. This section focuses on the materials that
helped us gain a better understanding of the image III. SYSTEM ARCHITECTURE
processing techniques research field's ideology.
The research articles went into great detail about
how different algorithms and frameworks can be
used for detection.

The authors in [3] developed a computer vision-

based method that makes use of Deep Convolutional
Neural Networks (DCNN) to find open spaces in any
parking lot. Although the method is beneficial for
locating parking spaces, it lacks any accompanying
structure that may aid drivers in finding parking
spaces at parking area entrances. Figure 1. System Architecture
In [4], the authors suggested a cloud-based design. IV. DESIGN METHOLOGY
The suggested architecture was implemented by the
authors using an Infrared sensor, a Raspberry Pi, a This research study has proposed a system that
cloud server, and a mobile application. The way the includes image processing techniques and OpenCV,
system is designed, the Raspberry Pi is connected to which is a most powerful computer vision library are
twelve sensors. Through the Web API, the device is used to find the number of vacant spaces. A fixed
registered so that it can be regarded as a component camera is used and videos are taken from the top
of the system. The sensor number allows for the view of the parking area. At the time of installation,
detection of the slot's value. The raw and a stable CCTV camera captured an image that
unprocessed data from Web APIs and Web Jobs are serves as the background reference image. There
sent to the Microsoft Azure SQL database via the are no automobiles in this reference picture. Frames
Microsoft Azure Web service. are used to segment video. Then, to reduce the
computational complexity, a key frame is extracted
In the technique shown in [5], the cameras capture from each segment and that segments are stored
the pictures and then extract their individual using pickle module in OpenCV. In image
attributes. This particular parking space is regarded processing, the pickle module can be used to save
as occupied if something has been placed in it that and load image data, such as pixel arrays or other
resembles an automobile. The obtained image is image features that have been extracted for use in
subjected to the image segmentation technique, after machine learning models. The segments are then
which the image is improved and the parking slot is subjected to additional processing. Finding the
found. parking spaces in the image is the main goal. The
In [6], LED lighting was regulated in accordance camera that is used to take the pictures is fixed in
with the detection outcome after a vehicle's presence place and always faces in the same general direction.
in the parking place was determined using infrared A high-definition camera is used to take a picture of
and ultrasonic sensors. By delivering the discovered a parking lot with cars in it. Lanes are separated

979-8-3503-4757-9/23/$31.00 ©2023 IEEE 1063

inside the visual frame that contains lanes. The the empty space, and a function is built to display the
obtained RGB image is first transformed to a vacant spaces with a green rectangular mark and the
grayscale image, and in the Image Segmentation occupied spaces with red. The number of cars is then
module, a binary image is then produced We have calculated and displayed as an output.
used an algorithm named GLCM(Gray-level- V. IMPLEMENTATION & RESULTS
cooccurrence-matrix) for feature extraction. It
Our proposed System detects the vacant spaces and
examines the relationship among pixels and defines
using image processing techniques and displays the
how frequently a combination of pixels are present
no. of vacant spaces available to park the car and
in a image in a given direction. For image
the total no. of parking slots available in that
enhancement we use kernel to highlight each pixel
parking area as an output.
and enhance the color it emits.

Using the adaptive thresholding technique, a binary

picture is created from the resulting grayscale image.
In adaptive thresholding, the threshold value is
computed based on local characteristics within
blocks of pixels, rather than using a single global
threshold value for the entire image. The threshold
level is chosen so that the important objects appear
white and the background is black. The position and
shape of interest are fully described by the binary
image.

Figure 5. Detection of vacant parking spaces

The Parking slot detection system has been

successful in showing whether a parking space is
occupied or not. Before the program was executed,
rectangle shape zones were set as input. The pixel
of the area will change if a vehicle enters the
designated zone of the specified Rectangle shape.
The program would then change the colour from
Figure 3. Image after Adaptive Thresholding
green to red as a result. The designated boxes are
shown in green when there is open pavement or a
The Weiner filter and other morphological processes
parking place that can be accessed, resulting in low
and dilation are used to reduce the amount of noise
in the binary image. pixels. The arrangement immediately changes from
green to red when a certain parking space is
occupied by a car since the pixel of that area has
increased.

VI. CONCLUSION

Numerous automatic parking systems using GSM,

wireless transmitter, and other technologies are
already in use. Being one of the most significant
technologies of our day and having a wide range of
uses, image processing was specifically picked for
this project with the intention of learning more about
it. The image processing-based parking space
Figure 4. Dilated Image recognition system in was created and put to the test.
With just a few CCTV, a big region may be
The vacant spaces are identified using the pixel managed. Due to the fact that it uses actual car
difference between the image containing the car and photos, it consistently detects approaching cars. Due

979-8-3503-4757-9/23/$31.00 ©2023 IEEE 1064

to the straightforward technology, it is inexpensive

A. Ampuni, S. Fonataba, A. Fitrianto and G. Wang, "Smart Parking
and simple to install. Through the guidance System With Automatic Cashier Machine Utilize the IoT
information display, drivers can obtain vital real- Technology," 2019
time parking lot information from this system. [9] J. Šilar, J. Růžička, Z. Bělinovà, M. Langr and K.
Hlubučková, "Smart parking in the smart city
application," 2018.

VII. FUTURE SCOPE [10] N. Jusat, A. A. Zainuddin, R. Sahak, A. B. Andrew, K.

Subramaniam and N. A. Rahman, "Critical Review In
The future study of this project can be expanded in Smart Car Parking Management Systems," 2021.
the future by including additional features like
automatic number plate recognition using image [11] P. Melnyk, S. Djahel and F. Nait-Abdesselam, "Towards a
Smart Parking Management System for Smart
processing, automated billing for parking the car, Cities," 2019 IEEE International Smart Cities
and guiding the user to the closest open parking Conference (ISC2), Casablanca, Morocco, 2019.
spaces and can concentrate on giving clients who
have already registered with an online parking [12] N. Farooqi, S. Alshehri, S. Nollily, L. Najmi, G. Alqurashi
and A. Alrashedi, "UParking: Developing a Smart
management system a specified spot. These features
Parking Management System Using the Internet of
could be combined with this system to create a Things," 2019
comprehensive intelligent system.
[13] N. ELKHALIDI, F. BENABBOU and N. SAEL, "Survey
of Reservation Techniques in Smart Parking," 2019.

REFERENCES [14] G. J. Leelipushpam Paulraj, I. Johnraja Jebadurai, P. F.

Khan, N. Sumanth and K. V. Reddy, "Smart Distributed
[1] Abhirup Khanna and Rishi Anand, "IoT based smart Parking System using Internet of Things," 2022 .
parking system", 2016 international conference of IoT
and applications.

[2] Sagar Rane, Aman Dubey and Tejisman Parida, "Design of

IoT Based Intelligent Parking System Using Image
Processing Algorithms",2017 Methodologies and
Communication.

[3] L. Zhang, J. Huang, X. Li, and L. Xiong, ‘‘Vision-based

parking-slot detection: A DCNN- based approach and a
large-scale benchmark dataset,’’IEEE Trans. Image
Process., vol. 27, no. 11, pp. 5350–5364, Nov. 2018.