IRJET- Automatic Number Plate Recognition using Image Segmentation
1 like39 views
This document describes research on automatic number plate recognition using image segmentation. The proposed methodology involves capturing an image of a vehicle, converting it to grayscale, performing thresholding for binarization, detecting the number plate region, segmenting individual characters, and recognizing the characters using optical character recognition. Previous literature on number plate recognition methods is reviewed, including works involving Persian, English, and Indian license plates. The proposed system aims to localize the plate region with 99% accuracy, segment characters with 99.2% accuracy, and recognize the plate numbers with 98% accuracy.
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 05 Issue: 08 | Aug 2018 www.irjet.net p-ISSN: 2395-0072
© 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 1607
AUTOMATIC NUMBER PLATE RECOGNITION USING IMAGE
SEGMENTATION
Ms. V. Poornima1, Ms. A. Carolin flora2, Mrs. J. Priyadharshini3
1,2Dept of computer science, Research Scholors, St.Joseph’s college for women- Tirupur, India.
3 Head of the Dept of computer science (PG), St.Joseph’s college for women- Tirupur, India.
--------------------------------------------------------------------***-----------------------------------------------------------------------
Abstract- This method of Automatic Number Plate
Recognition (ANPR) is one of the solutions of such a kind
of problem. There are various methodologies used but it is
a challenging task as some of the features like high speed
of vehicles, languages of number plate & mainly non-
uniform letter on number plate issues a lot in recognition.
License Plate Recognition plays main role on the traffic
monitoring and parking administration. They investigate
the vehicles and capture the image and the number plate
of vehicles is extracted from image using image
Segmentation and Optical Character Recognition
technique. The follow-on data compares the database
record so that we come up with the License plate number
such as open to industrial system which effectively detects
& recognizes the vehicle number plate on true image fair
when the pixel is of low resolution.
Keywords: Number plate recognition, Optical character
recognition, image segmentation, template matching.
1. INTRODUCTION
The ANPR (Automatic Number Plate
Recognition) plays a main part in various systems like
traffic observance system, Fault detection system, lifted
vehicle detection etc. Hence ANPR is used by the city
traffic department to monitor the traffic and track the
lifted vehicle. While ANPR is an extremely old research
area in image processing it is expanded year by year.
Identifying the number plate from the image or video is
not an easy task. The task of number plate writing style
varies from country to country so that recognizing the
number plate is difficult. In Indian number plate, writing
style changes from state to state. In India, the number
plate is distinct for two wheelers and four wheelers. For
Four wheelers, the number plate’s framework are also
different, i.e. yellow for tourist vehicles and white for
private cars. These are the basic challenges to keep in
mind before executing the ANPR system.
1.1 Image Capture: In this step, video image has to be
captured by any quality camera or by extracting the
attracted frame a stream of video. Capture the image
from the video stream and it’s needed an additional
work.
1.2 Image Preprocessing: Once the involved image is
being captured in which number plate obvious and fine
texture pattern, then the furthest processing of the
image is carried out. It has many steps: resize the image
resolution, removal of noise from an image, and
conversion of the image from RGB to gray and then
Binary (black and white).
1.3 Character Segmentation: After Preprocessing the
number plate region of the image is extract.
1.4 Optical Character Recognition (OCR): Electronic
conversion of handwritten or printed text images into
machine-encoded text. Here, OCR used to recognize the
number from the segmented Image.
Fig-1: Flow Diagram
2. LITERATURE REVIEW
In [1], it presents an online corrected structure
for automatic number plate recognition which can be
utilized as a reason for some genuine universes ITS
benefits. The framework is calculated to handle misty
vehicle number plates, varieties of climate and various
lighting conditions, distinctive activity circumstances,
and rapid vehicle plates. Additionally, it addresses
variable problems by indicate legitimate equipment
stages alongside ongoing, strong, and imaginative
calculations. And have assembled enormous and
unexpectedly comprehensive information sets of Persian
tags for assessments, examinations, and change of
different included calculations. The information sets
Capture the image from video
Processing of capture image
Character segmentation
Optical character recognition](https://image.slidesharecdn.com/irjet-v5i8276-180921063632/85/IRJET-Automatic-Number-Plate-Recognition-using-Image-Segmentation-1-320.jpg)
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 05 Issue: 08 | Aug 2018 www.irjet.net p-ISSN: 2395-0072
© 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 1608
include pictures which were fixed in the junction, roads,
roadways, day and night, different climate variations, a
particular number plate clarities. Using these
information sets, the structure accomplish 99% 99.2%,
and 98% exactness's to plate localization, character
division, and plate identification, separately. The
negative concern rate in plate localization is under 0.7%.
The general accuracy on the confused plate's
segment of our information sets is 91.5% The ANPR
framework has been introduced in a few areas and has
tried usually for over a year. The projected calculations
for every part of the framework are exceptionally fine to
lighting variations, measure varieties, number plate
clarity, and number plate imbalanced. The framework is
also autonomous of the quantity of number plates in
acquire pictures. The framework has been also tried on
three other the Iranian information sets and has
accomplished 100% exactness in both localization and
identification parts.
To demonstrate that the ANPI is not dialect
lower, they have tried the framework on accessible
English number plates information set and accomplished
97% general exactness. In [2], ANPR is built-up an
observation technique that uses optical character
identification on pictures to analyze the license plates on
vehicles. This framework is outlined with a neural
network which is prepared to recognize characters that
can be found in an Indian standard High sanctuary
Number Plate and is executed utilizing MATLAB. A
straightforward and effective framework has been
created to control the license plates from the picture of a
caught vehicle containing Indian criterion license plate.
A neural network builds a character identification
framework has been actualized to distinguish every one
of the characters that can be found in an Indian criterion
number plate.
A framework is set up to have great execution in
completing and coordinating the check example and
before put apart examples. Framework is talented and
gives good-looking result if there should be a happening
of slight variety in similar characters because of
disorder. The framework limit was experiential to be
0.85 which can be other enhanced preparing. For, same
character recognition remainder is as high as 0.937
while for various characters it is lesser than 0.5
accordingly the framework gives fine clearing up if there
should be a happening of various characters.
In [3], they execute RGB color extractor on different
sorts of tags. More than 225 color pictures taken by the
iPhone 5s camera are utilized as a part of this
examination. The test pictures are fixed from the front
and back of the vehicles below different conditions, for
example, unique edges, diverse luminance, and
individual climate conditions. Even though the fact that
the calculations were enhanced for the Illinois number
plate, which can be naturally enlarge out to perceive
other state tags of different situation of the United States.
RGB color observer is a standard instrument in picture
examination that allows us to separate the color data for
the pre-preparing in this procedure. The examining
outcomes display that the planned approach is about
forceful and practical. It maybe, there is opportunity to
get better in calculation because it doesn't work feasible
in situation below dark lights and mistakes as of various
states of characters we remove. The execution of analyze
tags from different states is additionally very much
fulfilled the achievement rate is near 100% which
indicate this strategy is rather skilled and exact at
extricating the characters with an empowering result.
The outcome examination of the framework gives 95.1%
exactness.
In [4], Automatic Number Plate Identification (ANPI)
framework screens and finds countless enlistment
license plates by perusing the vehicle license plates as
information and perceives the license plates' characters
as yield naturally. Truth is told, mistake of recognition
can be brought about by different variables, for example,
pivot of the plate and non-uniform light amid picture
procurement. In [4], de-skewing operation and format
coordinating procedure are planned to keep up the
precision of the auto license plate at the abnormal state.
All the information images needs to experience 5 phases
of improvement as wants to be, which incorporate pre-
handling stage, plate constraint organize, skew detection
and amendment arrange, character segmentation
organize and finally, character recognition organize for
the framework to create yield.
Each of the stages consists of demanding
frameworks that were tried and connected to complete
the ideal yield. At long last, it is to be turned out to be
100% precise for the plate localization, 99.6% for
character division, 91.5% for character recognition and
the general exactness of the framework is 91.1%. In [5],
one more approach is being presented for quick and
talented execution of ANPI framework.
In this move towards, the vertical boundary discovery
calculation is connected and evacuates undesirable
boundaries by picture equality method. The License
plate area is separated by unification factual and the
morphological image preparing methods. For character
identification, the layout coordinating is utilized for
optical character recognition (OCR). It functions
admirably in different regular situations self-
determining to varieties of shading, sort and size. This
approach can maybe work in most negative scenario
situations. A few adjustments in OCR are necessary for
recognition of a deeply partial font. The algorithm is
tried on 500 continuous pictures, which are procured
under various light location and various situations.
General production of the planned strategy is 84.8% and
the implementation time is under 0.5sec.](https://image.slidesharecdn.com/irjet-v5i8276-180921063632/85/IRJET-Automatic-Number-Plate-Recognition-using-Image-Segmentation-2-320.jpg)
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 05 Issue: 08 | Aug 2018 www.irjet.net p-ISSN: 2395-0072
© 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 1609
3. PROPOSED METHODOLOGY
Fig-2: STEPS INVOVLED IN VECHICLE NUMBER PLATE
RECONGITION
An Input image is captured by the camera.
Input image will be change to gray scale value. Then,
gray scale is change into double image by thresholding
method.
After, that we have to identify the size of the number
plate. But in general the plates are rectangle in shape
hence the edges of the plate are detected. Then, the
detection techniques are applied to calculate the
properties of the image region. Directly, after the type
connected components, the region will be exact from
the input image.
Now segmentation methods are applied to get
individual character and number image.
Finally, recognition techniques are applied for
identification of alienated characters and numbers.
Fig -3: Car license plate
4. CONCLUSION
In this paper various Number Plate recognition
strategies have been inspected in indirect elements
which were handling by several researchers. The
Number Plate Recognition (ANPR) framework is mostly
includes three significant strides, number plate
localization, character division and character
identification. Also, use of various methods and
techniques which are proposed by researchers earlier
are discussed. Personally have even mentioned the
basic and common steps involved in the vehicle number
plate identification. These types show the absolute in
order regarding how the traffic inspection systems used
the image processing methods and analysis tools for
detect, segment, and track the vehicles.
ACKNOWLEDGEMENT
I express my sincere gratitude towards my guide Prof.
Mrs. J. Priyadharshini M.C.A., M.Phil.,SET (Ph.D). for her
valuable guidance. I also thank Principal. Ref.sis.Dr. A.
Kulandai Therese M.Sc., M.Phil.,Ph.D., for their
encouragement and support. Their insight and
comments will definitely lead to a better presentation for
the ideas expressed in this paper.
REFERENCES
[1] Panahi, Rahim, and Iman Gholampour. "Accurate
Detection and Recognition of Dirty Vehicle Plate
Numbers for High-Speed Applications." IEEE
Transactions on Intelligent Transportation Systems
(2016).
[2] Fajas, F., et al. "Automatic Number Plate Recognition
for Indian standard number plates." Ultra Modern
Telecommunications and Control Systems and
Workshops (ICUMT), 2012 4th International Congress
on. IEEE, 2012.
[3] Jia, Yonghui, Thomas Gonnot, and Jafar Saniie.
"Design flow of vehicle License Plate reader based on
RGB color extractor." Electro Information Technology
(EIT), 2016 IEEE International Conference on. IEEE,
2016.
[4] Keong, Wong Weng, and Vahab Iranmanesh.
"Malaysian automatic number plate recognition
framework using Pearson correlation." Computer
Applications & Industrial Electronics (ISCAIE), 2016
IEEE Symposium on. IEEE, 2016.
[5] Saleem, Nauman, et al. "Automatic license plate
recognition using extracted features." Computational
and Business Intelligence (ISCBI), 2016 4th International
Symposium on. IEEE, 2016.](https://image.slidesharecdn.com/irjet-v5i8276-180921063632/85/IRJET-Automatic-Number-Plate-Recognition-using-Image-Segmentation-3-320.jpg)
IRJET- Automatic Number Plate Recognition using Image Segmentation
- 1. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 05 Issue: 08 | Aug 2018 www.irjet.net p-ISSN: 2395-0072 © 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 1607 AUTOMATIC NUMBER PLATE RECOGNITION USING IMAGE SEGMENTATION Ms. V. Poornima1, Ms. A. Carolin flora2, Mrs. J. Priyadharshini3 1,2Dept of computer science, Research Scholors, St.Joseph’s college for women- Tirupur, India. 3 Head of the Dept of computer science (PG), St.Joseph’s college for women- Tirupur, India. --------------------------------------------------------------------***----------------------------------------------------------------------- Abstract- This method of Automatic Number Plate Recognition (ANPR) is one of the solutions of such a kind of problem. There are various methodologies used but it is a challenging task as some of the features like high speed of vehicles, languages of number plate & mainly non- uniform letter on number plate issues a lot in recognition. License Plate Recognition plays main role on the traffic monitoring and parking administration. They investigate the vehicles and capture the image and the number plate of vehicles is extracted from image using image Segmentation and Optical Character Recognition technique. The follow-on data compares the database record so that we come up with the License plate number such as open to industrial system which effectively detects & recognizes the vehicle number plate on true image fair when the pixel is of low resolution. Keywords: Number plate recognition, Optical character recognition, image segmentation, template matching. 1. INTRODUCTION The ANPR (Automatic Number Plate Recognition) plays a main part in various systems like traffic observance system, Fault detection system, lifted vehicle detection etc. Hence ANPR is used by the city traffic department to monitor the traffic and track the lifted vehicle. While ANPR is an extremely old research area in image processing it is expanded year by year. Identifying the number plate from the image or video is not an easy task. The task of number plate writing style varies from country to country so that recognizing the number plate is difficult. In Indian number plate, writing style changes from state to state. In India, the number plate is distinct for two wheelers and four wheelers. For Four wheelers, the number plate’s framework are also different, i.e. yellow for tourist vehicles and white for private cars. These are the basic challenges to keep in mind before executing the ANPR system. 1.1 Image Capture: In this step, video image has to be captured by any quality camera or by extracting the attracted frame a stream of video. Capture the image from the video stream and it’s needed an additional work. 1.2 Image Preprocessing: Once the involved image is being captured in which number plate obvious and fine texture pattern, then the furthest processing of the image is carried out. It has many steps: resize the image resolution, removal of noise from an image, and conversion of the image from RGB to gray and then Binary (black and white). 1.3 Character Segmentation: After Preprocessing the number plate region of the image is extract. 1.4 Optical Character Recognition (OCR): Electronic conversion of handwritten or printed text images into machine-encoded text. Here, OCR used to recognize the number from the segmented Image. Fig-1: Flow Diagram 2. LITERATURE REVIEW In [1], it presents an online corrected structure for automatic number plate recognition which can be utilized as a reason for some genuine universes ITS benefits. The framework is calculated to handle misty vehicle number plates, varieties of climate and various lighting conditions, distinctive activity circumstances, and rapid vehicle plates. Additionally, it addresses variable problems by indicate legitimate equipment stages alongside ongoing, strong, and imaginative calculations. And have assembled enormous and unexpectedly comprehensive information sets of Persian tags for assessments, examinations, and change of different included calculations. The information sets Capture the image from video Processing of capture image Character segmentation Optical character recognition
- 2. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 05 Issue: 08 | Aug 2018 www.irjet.net p-ISSN: 2395-0072 © 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 1608 include pictures which were fixed in the junction, roads, roadways, day and night, different climate variations, a particular number plate clarities. Using these information sets, the structure accomplish 99% 99.2%, and 98% exactness's to plate localization, character division, and plate identification, separately. The negative concern rate in plate localization is under 0.7%. The general accuracy on the confused plate's segment of our information sets is 91.5% The ANPR framework has been introduced in a few areas and has tried usually for over a year. The projected calculations for every part of the framework are exceptionally fine to lighting variations, measure varieties, number plate clarity, and number plate imbalanced. The framework is also autonomous of the quantity of number plates in acquire pictures. The framework has been also tried on three other the Iranian information sets and has accomplished 100% exactness in both localization and identification parts. To demonstrate that the ANPI is not dialect lower, they have tried the framework on accessible English number plates information set and accomplished 97% general exactness. In [2], ANPR is built-up an observation technique that uses optical character identification on pictures to analyze the license plates on vehicles. This framework is outlined with a neural network which is prepared to recognize characters that can be found in an Indian standard High sanctuary Number Plate and is executed utilizing MATLAB. A straightforward and effective framework has been created to control the license plates from the picture of a caught vehicle containing Indian criterion license plate. A neural network builds a character identification framework has been actualized to distinguish every one of the characters that can be found in an Indian criterion number plate. A framework is set up to have great execution in completing and coordinating the check example and before put apart examples. Framework is talented and gives good-looking result if there should be a happening of slight variety in similar characters because of disorder. The framework limit was experiential to be 0.85 which can be other enhanced preparing. For, same character recognition remainder is as high as 0.937 while for various characters it is lesser than 0.5 accordingly the framework gives fine clearing up if there should be a happening of various characters. In [3], they execute RGB color extractor on different sorts of tags. More than 225 color pictures taken by the iPhone 5s camera are utilized as a part of this examination. The test pictures are fixed from the front and back of the vehicles below different conditions, for example, unique edges, diverse luminance, and individual climate conditions. Even though the fact that the calculations were enhanced for the Illinois number plate, which can be naturally enlarge out to perceive other state tags of different situation of the United States. RGB color observer is a standard instrument in picture examination that allows us to separate the color data for the pre-preparing in this procedure. The examining outcomes display that the planned approach is about forceful and practical. It maybe, there is opportunity to get better in calculation because it doesn't work feasible in situation below dark lights and mistakes as of various states of characters we remove. The execution of analyze tags from different states is additionally very much fulfilled the achievement rate is near 100% which indicate this strategy is rather skilled and exact at extricating the characters with an empowering result. The outcome examination of the framework gives 95.1% exactness. In [4], Automatic Number Plate Identification (ANPI) framework screens and finds countless enlistment license plates by perusing the vehicle license plates as information and perceives the license plates' characters as yield naturally. Truth is told, mistake of recognition can be brought about by different variables, for example, pivot of the plate and non-uniform light amid picture procurement. In [4], de-skewing operation and format coordinating procedure are planned to keep up the precision of the auto license plate at the abnormal state. All the information images needs to experience 5 phases of improvement as wants to be, which incorporate pre- handling stage, plate constraint organize, skew detection and amendment arrange, character segmentation organize and finally, character recognition organize for the framework to create yield. Each of the stages consists of demanding frameworks that were tried and connected to complete the ideal yield. At long last, it is to be turned out to be 100% precise for the plate localization, 99.6% for character division, 91.5% for character recognition and the general exactness of the framework is 91.1%. In [5], one more approach is being presented for quick and talented execution of ANPI framework. In this move towards, the vertical boundary discovery calculation is connected and evacuates undesirable boundaries by picture equality method. The License plate area is separated by unification factual and the morphological image preparing methods. For character identification, the layout coordinating is utilized for optical character recognition (OCR). It functions admirably in different regular situations self- determining to varieties of shading, sort and size. This approach can maybe work in most negative scenario situations. A few adjustments in OCR are necessary for recognition of a deeply partial font. The algorithm is tried on 500 continuous pictures, which are procured under various light location and various situations. General production of the planned strategy is 84.8% and the implementation time is under 0.5sec.
- 3. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 05 Issue: 08 | Aug 2018 www.irjet.net p-ISSN: 2395-0072 © 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 1609 3. PROPOSED METHODOLOGY Fig-2: STEPS INVOVLED IN VECHICLE NUMBER PLATE RECONGITION An Input image is captured by the camera. Input image will be change to gray scale value. Then, gray scale is change into double image by thresholding method. After, that we have to identify the size of the number plate. But in general the plates are rectangle in shape hence the edges of the plate are detected. Then, the detection techniques are applied to calculate the properties of the image region. Directly, after the type connected components, the region will be exact from the input image. Now segmentation methods are applied to get individual character and number image. Finally, recognition techniques are applied for identification of alienated characters and numbers. Fig -3: Car license plate 4. CONCLUSION In this paper various Number Plate recognition strategies have been inspected in indirect elements which were handling by several researchers. The Number Plate Recognition (ANPR) framework is mostly includes three significant strides, number plate localization, character division and character identification. Also, use of various methods and techniques which are proposed by researchers earlier are discussed. Personally have even mentioned the basic and common steps involved in the vehicle number plate identification. These types show the absolute in order regarding how the traffic inspection systems used the image processing methods and analysis tools for detect, segment, and track the vehicles. ACKNOWLEDGEMENT I express my sincere gratitude towards my guide Prof. Mrs. J. Priyadharshini M.C.A., M.Phil.,SET (Ph.D). for her valuable guidance. I also thank Principal. Ref.sis.Dr. A. Kulandai Therese M.Sc., M.Phil.,Ph.D., for their encouragement and support. Their insight and comments will definitely lead to a better presentation for the ideas expressed in this paper. REFERENCES [1] Panahi, Rahim, and Iman Gholampour. "Accurate Detection and Recognition of Dirty Vehicle Plate Numbers for High-Speed Applications." IEEE Transactions on Intelligent Transportation Systems (2016). [2] Fajas, F., et al. "Automatic Number Plate Recognition for Indian standard number plates." Ultra Modern Telecommunications and Control Systems and Workshops (ICUMT), 2012 4th International Congress on. IEEE, 2012. [3] Jia, Yonghui, Thomas Gonnot, and Jafar Saniie. "Design flow of vehicle License Plate reader based on RGB color extractor." Electro Information Technology (EIT), 2016 IEEE International Conference on. IEEE, 2016. [4] Keong, Wong Weng, and Vahab Iranmanesh. "Malaysian automatic number plate recognition framework using Pearson correlation." Computer Applications & Industrial Electronics (ISCAIE), 2016 IEEE Symposium on. IEEE, 2016. [5] Saleem, Nauman, et al. "Automatic license plate recognition using extracted features." Computational and Business Intelligence (ISCBI), 2016 4th International Symposium on. IEEE, 2016.