Contenu connexe
Similaire à 20120130402031
Similaire à 20120130402031 (20)
Plus de IAEME Publication
Plus de IAEME Publication (20)
20120130402031
- 1. International Journal of Advanced Research in Engineering and Technology (IJARET), ISSN 0976 –
6480(Print), ISSN 0976 – 6499(Online) Volume 4, Issue 2, March – April (2013), © IAEME
295
LICENSE PLATE RECOGNITION SYSTEM FOR INDIAN VEHICLES
M. M. Kodabagi1
, Mr. Vijayamahantesh S. Kanavi2
1
Department of Computer Science and Engineering, Basaveshwar Engineering College,
Bagalkot-587102, Karnataka, India
2
Department of Computer Science & Engineering, K.L.E. Institute of Technology,
Hubli-580030, Karnataka, India
ABSTRACT
License Plate Recognition (LPR) has been intensively studied in many countries. Due
to the different types of number plates being used, the requirements of an automatic number
plate recognition system are different for each country. In general, objective of any LPR
system is to localize and recognize potential license plate region(s) of the vehicle images
captured from handheld device/digital camera/mobile phone camera. The proposed method
for License Plate Recognition (LPR) system works in three modules: localization of license
plate, segmentation of the characters and recognition of the characters from the license plate.
Localization of the license plate is done using morphological operations, horizontal &
vertical edge processing. Character segmentation is carried out using connected component
labeling. Character recognition is done by using Neural Network classifier. The method is
tested on 100 samples of Indian vehicle images. The system achieves 86% accuracy in
localization, 81% accuracy in segmentation and 80% accuracy in character recognition.
Keywords: Localization, Segmentation, Recognition, Neural Network
1. INTRODUCTION
During the past few years, Intelligent Transportation Systems (ITSs) have had a wide
impact in people’s life as their scope is to improve transportation safety and mobility and to
enhance productivity through the use of advanced technologies. In the current information
technology era, the use of automations and intelligent systems is becoming more and more
widespread. The Intelligent Transport System (ITS) technology has received so much
attention that many systems are being developed and applied all over the world. Therefore
License Plate Recognition (LPR) has turned out to be an important research issue.
INTERNATIONAL JOURNAL OF ADVANCED RESEARCH IN
ENGINEERING AND TECHNOLOGY (IJARET)
ISSN 0976 - 6480 (Print)
ISSN 0976 - 6499 (Online)
Volume 4, Issue 2 March – April 2013, pp. 295-304
© IAEME: www.iaeme.com/ijaret.asp
Journal Impact Factor (2013): 5.8376 (Calculated by GISI)
www.jifactor.com
IJARET
© I A E M E
- 2. International Journal of Advanced Research in Engineering and Technology (IJARET), ISSN 0976 –
6480(Print), ISSN 0976 – 6499(Online) Volume 4, Issue 2, March – April (2013), © IAEME
296
License Plate Recognition has many applications in traffic monitoring system,
including controlling the traffic volume, ticketing vehicle without the human control, vehicle
tracking, policing, security, and so on. The most vital and the most difficult part of any LPR
system is the detection and extraction of the vehicle number plate, which directly affects the
systems overall accuracy. The presence of noise, blurring in the image, uneven illumination,
dim light, variation in license plate sizes and foggy conditions make the task even more
difficult [1]. Therefore the overall problem may be subdivided into three distinct key
modules: (a) localization of license plate from vehicle image (b) segmentation of the
characters within the license plate and (c) recognition of segmented characters within
the license plate [4].
The main function of the module (a) is to find out the potential regions within
the image that may contain the license plate. The function of module (b) is to isolate
the foreground characters from the background within the detected license plate region.
And the function of the module (c) is to recognize the segments in terms of known characters
or digits. Though modules (b) and (c) employ most of the traditional methods available
to the technologists, module (a) i.e. localization of potential license plate regions(s)
from vehicle images is the most challenging task due to the huge variations in size,
shape, color, texture and spatial orientations of license plate regions in such images.
Hence, in this work a novel method is proposed for detecting the location of vehicle number
plates and recognizing the numerals and characters in it.
In this paper, a new model for License Plate Recognition (LPR) system is presented.
The proposed method for License Plate Recognition (LPR) system works in three modules:
localization of license plate, segmentation of the characters and recognition of the characters
from the license plate. Localization of the license plate is done using morphological
operations, horizontal & vertical edge processing. Character segmentation is carried out using
connected component labeling. Character recognition is done by using Neural Network
classifier. The method is tested on 100 samples of Indian vehicle images. The system
achieves 86% accuracy in localization, 81% accuracy in segmentation and 80% accuracy in
character recognition.
The rest of the paper is organized as follows: The related work is discussed in section
II. The proposed model is presented in section III. Experimental results are given in section
IV. Section V concludes the work and lists future directions.
2. RELATED WORKS
A substantial amount of work has gone into the research related to License Plate
Recognition system. Some of the related works are summarized below.
An approach for Automatic Number Plate extraction, character segmentation and
recognition for Indian vehicles is proposed in [1]. In India, number plate models are not
followed strictly. Characters on plate are in different Indian languages, as well as in English.
Due to variations in the representation of number plates, vehicle number plate extraction,
character segmentation and recognition are crucial. We present the number plate extraction,
character segmentation and recognition work, with English characters. Number plate
extraction is done using Sobel filter, morphological operations and connected component
analysis. Character segmentation is done by using connected component and vertical
projection analysis. Character recognition is carried out using Support Vector Machine
(SVM). The segmentation accuracy is 80% and recognition rate is 79.84 %.
- 3. International Journal of Advanced Research in Engineering and Technology (IJARET), ISSN 0976 –
6480(Print), ISSN 0976 – 6499(Online) Volume 4, Issue 2, March – April (2013), © IAEME
297
A new technique for Automatic Number Plate Recognition is discussed in [2]. This
new approach use the vehicle number plate image captured from the digital camera and the
image is processed to get the number plate information. In this context, the number plate area
is localized using a novel Feature-based number plate localization method mainly focusing on
the two algorithms i.e., Edge Finding Method and Window Filtering Method. These methods
are used for the better development of the number plate detection system.
A new method for pixel based segmentation algorithm of the alphanumeric characters
in the license plate is presented in [3]. This system can be used in many different countries –a
feature which can be especially useful for trans-border traffic e.g. use in country borders etc.
Additionally, there is an option available to the end-user for retraining the Artificial Neural
Network (ANN) by building a new sample font database. The system was tested on 150
different number plates from various countries and an accuracy of 91.59% has been reached.
Another method for License Plate Recognition system is presented in [4]. This new
method used an innovative idea statistical distribution of the vertical edges for localization of
license plate, connected component labeling for segmentation of the characters and template
matching technique for recognition of the characters. In this method a real time vehicle
images are captured from a road-side surveillance camera automatically throughout day and
night. The images are stored in a centralized data server. A continuous process takes the
stored images sequentially and interprets the license number of the vehicle. The performance
of the system is measured at the three levels, i.e., localization level, segmentation level and
recognition level and the result seems are quite satisfactory.
A robust method for license plate location, segmentation and recognition of the
characters proposed in [5]. The images of various vehicles have been acquired manually and
converted in to gray-scale images. Then wiener2 filter is used to remove noise present in the
plates. The segmentation of gray scale image generated by finding edges using Sobel filter
and then bwlabel is used to count the number of connected component. Finally, single
character is detected. The results show that the proposed method achieved accuracy of 98%
by optimizing various parameters with higher recognition rate than the traditional methods.
A new technique for real time application which recognizes license plates from cars at
a gate is employed in [6]. The system is based on regular PC with video camera, catches
video frames which include a visible car license plate and processes them. Once a license
plate is detected, its digits are recognized, displayed on the user interface or checked against a
database. The focus is on the design of algorithms used for extracting the license plate from a
single image, isolating the characters of the plate and identifying the individual characters.
The proposed system has been implemented using Vision Assistant 8.2.1 & Labview 11.0 the
recognition of about 98% vehicles shows that the system is quite efficient.
A new approach for the Automatic Recognition of Vehicle License Plates is presented
in [7]. In this approach Multilayer feed-forward back-propagation algorithm using three
hidden layers are used. The performance of the proposed algorithm has been tested on real
vehicle images. Based on the experimental results, they conclude that their algorithm shows
superior performance using two hidden layers than single hidden layer and the results
revealed that as the number of hidden layers is increased, more accuracy is achieved.
A new method for License Plate Recognition System proposed in [8]. This method
uses vehicle steps over magnetic loop detector that senses car and takes image of the car,
following image preprocessing operations for improvement in the quality of car image. From
this enhanced image, license plate region is recognized and extracted. Then character
fragmentation/segmentation is performed on extracted License Plate and those segmented
characters are recognized using Neural Network.
- 4. International Journal of Advanced Research in Engineering and Technology (IJARET), ISSN 0976 –
6480(Print), ISSN 0976 – 6499(Online) Volume 4, Issue 2, March – April (2013), © IAEME
298
An algorithm based on a combination of morphological operation with area criteria
tests for number plate localization is presented in [9]. In this method segmentation of the
number plate characters is achieved by the application of edge detectors, labeling and fill hole
approach. The character recognition was accomplished with the aid of optical characters by
the process of Template matching. The system was experimented with four different edge
detectors namely Sobel, Canny, Prewitt, LOG. A comparative analysis on the success rate of
the proposed system showed overall better success rate of 96.8% by using canny edge
detector.
An approach for Automatic Number Plate Recognition system discussed in [10]. This
approach consists of few algorithm like “Feature based number plate Localization” for
locating the number plate, “Image Scissoring” algorithm for character segmentation and
proposed algorithm for character recognition using Support Vector Machine (SVM).System
can recognize single or double line number plate.
The problem of license plate recognition using a three layer fuzzy neural network is
presented in [11]. In the first stage the plate is detected inside the digital image using
rectangular perimeter detection and the finding of a pattern by pattern matching, after that,
the characters are extracted from the plate by means of horizontal and vertical projections.
Finally, a fuzzy neural network is used to recognize the license plate. The tests were made in
an uncontrolled environment in a parking lot and using Mexican and American plates. The
results show that the system is robust as compare to those systems Papered in the literature.
A new method for Vehicle License Plate Identification is employed in [12]. In this
method an adaptive image segmentation technique (sliding concentric windows) and
connected component analysis in conjunction with a character recognition neural network.
The algorithm was tested with 1334 natural-scene gray-level vehicle images of different
backgrounds and ambient illumination. The camera focused in the plate, while the angle of
view and the distance from the vehicle varied ac-cording to the experimental setup. The
license plates properly segmented were 1287 over 1334 input images (96.5%). The optical
character recognition system is a two-layer probabilistic neural network (PNN) with topology
108-180-36, whose performance for entire plate recognition reached 89.1%. The PNN is
trained to identify alphanumeric characters from car license plates based on data obtained
from algorithmic image processing. Combining the above two rates, the overall rate of
success for the license-plate-recognition algorithm is 86.0%.
Out of many works cited in the literature it is agreed that existing algorithms based on
localization, segmentation and recognition of characters in vehicle license plates are having
some limitations due to, variations in the representation of number plates, color of the plate,
font/size of each character, spacing between subsequent characters etc. To address these
challenges the objective of current work is to develop a novel technique for localization,
segmentation and recognition of license plate not only to enrich the research but also used in
full-fledged commercial environment.
3. PROPOSED MODEL
The proposed model comprises in two main phases, Training and Testing. The
Training phase covers preprocessing of training images, zone-wise statistical feature
extraction, knowledge base construction. The Testing phase covers preprocessing of car
image, number plate localization, character segmentation and recognition model. The block
diagram of the proposed model is given in Fig. 1. The detailed description of each phase is
presented in the following subsections.
- 5. International Journal of Advanced Research in Engineering and Technology (IJARET), ISSN 0976 –
6480(Print), ISSN 0976 – 6499(Online) Volume 4, Issue 2, March – April (2013), © IAEME
299
.
Fig. 1. Block Diagram of Proposed Model
The images are collected from various vehicles captured from mobile phone camera
(2 MegaPixels). The characters from those license plates are segmented manually and stored
as training images/samples in the database for training neural network.
3.1 Preprocessing of training image
The character image from the database is preprocessed for enhancing the recognition
accuracy of the proposed work. The steps of preprocessing are as follows:
• Convert the input image into black and white image using the function im2bw.
• Salt and pepper noise are removed by calling the medfilt2 function.
• Image is resize to 30*30 by using the function imresize..
• Generate thinned image by calling the function bwmorph.
3.2 Feature Extraction
The preprocessed image is collected from database then it is divided into various
blocks/zones with the size 5 X 5 as shown in Fig. 2.2. The sum of all pixels in the zone is
computed which represent one feature/value of an image. Totally 36(30 * 30 size of the
image) features are computed in the same manner for an entire image and features are stored
in the feature vector X. The feature vector X is described in equation (1).
X = [ fi ]; 361 ≤≤ i (1)
Where,
fi is feature vector of ith
zone.
- 6. International Journal of Advanced Research in Engineering and Technology (IJARET), ISSN 0976 –
6480(Print), ISSN 0976 – 6499(Online) Volume 4, Issue 2, March – April (2013), © IAEME
300
3.3 Knowledge Base Construction.
From the database, 50% of samples are used for training and 50% of samples are used
for testing. Then the feature vectors of all training images are stored into knowledge base KB
as shown in equation (2).
KB = [ Xj ]; Nj ≤≤1 (2)
Where
• KB is knowledge base,
• Xj is a feature vector of jth
image in the training dataset
• N is the number of training samples.
3.4 Training the Neural Network
The Knowledge base is given as input to the Neural Network for training of samples.
3.5 Preprocessing of vehicle image
In this process Colored Image is converted into Gray-Scale Image. The algorithm
described here is independent of the type of colors in image and relies mainly on the
gray level of an image for processing and extracting the required information. Color
components like Red, Green and Blue value are not used throughout this algorithm. So, if
the input image is a colored image represented by 3-dimensional array, it is converted
to a 2-dimensional gray image before further processing.
3.6. Number Plate Localization
Dilation is a process of improvising given image by filling holes in an image,
sharpen the edges of objects in an image, and join the broken lines and increase the
brightness of an image. In this process, each pixel is compared with its neighboring pixels
and its value is set equal to the maximum value out of both the neighboring pixels. This
process makes edges of an image sharper. In turn, it helps in better detection of an image. If
an input image is blurred, this step will help to improve such blurred image and make it easy
for detection.
In horizontal edge processing of an image, the algorithm traverses through each
column of an image. In each column, the algorithm starts with the second pixel from the top.
The difference between second and first pixel is calculated. If the difference exceeds certain
threshold, it is added to total sum of differences. Then, algorithm will move downwards to
calculate the difference between the third and second pixels. So on, it moves until the end of
a column and calculate the total sum of differences between neighboring pixels. At the end,
an array containing the column-wise sum is created. The same process is carried out to find
the vertical edge processing. In this case, rows are processed instead of columns.
To prevent loss of important information in further steps, it is required to smooth
out such changes in values of histogram. For the same, the histogram value is passed
through a low-pass digital filter. While performing this step, each histogram value is
averaged, considering the values on it right-hand side and left-hand side. This step is
performed on both horizontal as well as the vertical array values.
This step finds all the regions in an image that has high probability of containing a
license plate. Co-ordinates of all such probable regions are stored in an array.
- 7. International Journal of Advanced Research in Engineering and Technology (IJARET), ISSN 0976 –
6480(Print), ISSN 0976 – 6499(Online) Volume 4, Issue 2, March – April (2013), © IAEME
301
The output of segmentation process is all the regions that have maximum probability
of containing a license plate. Out of these regions, the one with the maximum histogram
value is considered as the most probable candidate for number plate. All the regions
are processed row-wise and column-wise to find a common region having maximum
horizontal and vertical histogram value. This is the region having highest probability of
containing a license plate.
3.7 Character segmentation
In character segmentation module, global threshold (level) is computed that can be
used to convert an intensity image to a binary image. All pixels in the input image with
luminance greater than a level value with the value 1 (white) and replace all other pixels with
the value 0 (black).
Label connected components
Once we got the a threshold value the next step is to remove from a binary image all
connected components (objects) that have fewer than 30 pixels, producing another binary
image. The bwareaopen function used in this work for such purpose. The default connectivity
is 8 for two dimensions.
Measure properties of image regions
The next step is used to measure properties of image regions. The image is a logical
array; it can have any dimension. In grayscale image Bounding Box method is used, which
returns the smallest rectangle containing the region. The method takes two parameters, one to
specify the upper-left corner of the bounding box and another specifies the width of the
bounding box along each dimension. Once the properties are measured a rectangle box is
drawn on each of the object identified.
Objects extraction
In the object extraction, the find function returns the row and column indices of the
nonzero entries in the matrix X. This syntax is especially useful when working with sparse
matrices. Individual objects are extracted one by one and these characters are held in an array
for further processing.
3.8 Character Recognition Model
In this phase, a testing image is processed to obtain zone-wise features and stored into
a feature vector T as shown in equation (3).
T = [ fi ]; 361 ≤≤ i (3)
Character classifier in this system is based on K-Nearest Neighbors (KNN). The
design of KNN based character classifier is shown in figure 2.
- 8. International Journal of Advanced Research in Engineering and Technology (IJARET), ISSN 0976 –
6480(Print), ISSN 0976 – 6499(Online) Volume 4, Issue 2, March – April (2013), © IAEME
302
Figure 2 KNN-based Character Classifier
The KNN classifier uses features of testing image returns recognized character.
4. EXPERIMENTAL RESULTS AND ANALYSIS
The experimental tests were conducted for most of the images containing several
issues like spatial orientation, double lined characters, tilted license plates and results were
highly encouraging. The experimental result of processing one of the issue i.e license plate
with spatial orientation is considered as an example and the overall performance of the
system are given below;
Fig 3.Image with license plate at Fig 4. After Gray Scale conversion
spatial orientation
Fig 5 After localization phase Fig 6. After Segmentation phase
Fig 7. Recognised Characters
- 9. International Journal of Advanced Research in Engineering and Technology (IJARET), ISSN 0976 –
6480(Print), ISSN 0976 – 6499(Online) Volume 4, Issue 2, March – April (2013), © IAEME
303
TABLE 1: Overall System Performance
5. CONCLUSION
The work presented in this paper involved developing a License Plate Recognition
system in order to verify character recognition of the vehicle number plate. From the
experimental results it is observed that this method gives the good results. The method is
tested for 100 samples of vehicle image database. This technique is not only
computationally less expensive but also provides best recognition results. The method
is insensitive to the presence of skew, different font size/style, double lined number plate
and achieves 86% accuracy in localization,81% accuracy in segmentation and 80% accuracy
in character recognition accuracy. Several open questions still remain in this license plate
recognition system. The robustness for image variations in rotations, illumination, etc. must
be improved.
REFERENCES
[1] M. M. Shidore, S. P. Narote, ”Number Plate Recognition for Indian Vehicles”,
IJCSNS International Journal of Computer Science and Network Security, VOL.11
No.2, Feb. 2011, pp.143-146.
[2] S.Kranthi, K.Pranathi, A.Srisaila, “Automatic Number plate Recognition”,
International Journal of Advancements in Technology, Vol 2, No. 3, July 2011,
pp. 408-422.
[3] Ankush Roy, Debarshi Patanjali Ghoshal, “Number Plate Recognition for use in
Different Countries Using an Improved Segmentation”, 978-1-4244-9581-8/11/$26.00
© 2011 IEEE.
[4] Satadal Saha, Subhadip Basu and Mita Nasipuri, “Automatic Localization and
Recognition of License Plate Characters for Indian Vehicles”, IJCSET, Vol-2 No 4
August 2011, pp. 520-533.
[5] Anish Lazrus, Siddhartha Choubey, Sinha G.R.”An Efficient Method of Vehicle
Number Plate Detection And Recognition”, IJMI International Journal of Machine
Intelligence ISSN: 0975–2927 & E-ISSN: 0975–9166, Volume 3, Issue 3, 2011,
pp-134-137.
[6] Kuldeepak, Monika kaushik, Munish Vashishath, “License Plate Recognition System
based on Image Processing Using Labview”, International Journal of Electronics
Communication and Computer Technology (IJECCT), Volume 2 Issue 4 (July 2012),
pp. 183-188.
[7] Stuti Asthana, Niresh Sharma, Rajdeep Singh, “Vehicle number plate recognition
using multiple layer back propagation neural networks”, International Journal of
Computer Technology and Electronics Engineering (IJCTEE) Volume 1, Issue 1,
pp.35-38.
Total Number of
Images used for
Testing
Localization
Model
Segmentation
Model
Character Recognition
Model
100 86% 81% 80%
- 10. International Journal of Advanced Research in Engineering and Technology (IJARET), ISSN 0976 –
6480(Print), ISSN 0976 – 6499(Online) Volume 4, Issue 2, March – April (2013), © IAEME
304
[8] Anuja P. Nagare, “License Plate Character Recognition System using Neural
Network”, International Journal of Computer Applications (0975 – 8887), Volume
25– No.10, July 2011 pp.36-39.
[9] P.Anishiya, Prof. S. Mary Joans, “Number Plate Recognition for Indian Cars using
Morphological Dilation and Erosion with the Aid of Ocrs” , 2011 International
Conference on Information and Network Technology IACSIT Press, Singapore,
IPCSIT vol.4 2011, pp.116-119.
[10] Shriram Kishanrao Waghmare, A. K. Gulve, Vikas N.Nirgude, “Automatic Number
Plate Recognition (Anpr) System For Indian Conditions using Support Vector
Machine (Svm)”, International Journal of Computer Science and its Applications
[ISSN 2250 - 3765], pp.216-220.
[11] Osslan Osiris Vergara Villegas, Daniel Gonzál ez Balderrama, Humberto de Jesús
Ochoa Domínguez and Vianey Guadalupe Cruz Sánchez, ”License Plate Recognition
Using a Novel Fuzzy Multilayer Neural Network”, International Journal of
Computers, Issue 1, Volume 3, 2009, pp.31-40.
[12] Christos Nikolaos E. Anagnostopoulos, Ioannis E. Anagnostopoulos, Vassili Loumos,
and Eleftherios Kayafas, “A License Plate-Recognition Algorithm for Intelligent
Transportation System Applications”, IEEE, Transactions on Intelligent
Transportation Systems. Vol. 7, No.3, September 2006. pp. 377-392.
[13] Mayur V. Parulekar, Viraj Padte, Dhaval Shroff, Akash Metawala and Harsh
Nangalia, “Congestion Control & Collision Avoidance Algorithm in Intelligent
Transportation Systems”, International Journal of Computer Engineering &
Technology (IJCET), Volume 4, Issue 2, 2013, pp. 114 - 124, ISSN Print:
0976 – 6367, ISSN Online: 0976 – 6375.
[14] M. M. Kodabagi, S. A. Angadi and Chetana. R. Shivanagi, “Character Recognition of
Kannada Text in Scene Images using Neural Network”, International Journal of
Graphics and Multimedia (IJGM), Volume 4, Issue 1, 2013, pp. 9 - 19, ISSN Print:
0976 – 6448, ISSN Online: 0976 –6456.
[15] M. M. Kodabagi and S. R. Karjol, “Script Identification from Printed Document
Images using Statistical Features”, International Journal of Computer Engineering &
Technology (IJCET), Volume 4, Issue 2, 2013, pp. 607 - 622, ISSN Print:
0976 – 6367, ISSN Online: 0976 – 6375.
[16] M. M. Kodabagi and Hemavati C. Purad, “Script Identification using DCT
Coefficients”, International Journal of Graphics and Multimedia (IJGM), Volume 4,
Issue 1, 2013, pp. 31 - 40, ISSN Print: 0976 – 6448, ISSN Online: 0976 –6456.
[17] M. M. Kodabagi, S. A. Angadi and Anuradha. R. Pujari, “Text Region Extraction
from Low Resolution Display Board Images using Wavelet Features”, International
Journal of Information Technology and Management Information Systems (IJITMIS),
Volume 4, Issue 1, 2013, pp. 38 - 49, ISSN Print: 0976 – 6405, ISSN Online:
0976 – 6413.