Vehicle License Plate Location Based on Mathematical Morphology and Variance

Projection
Abstract- In a vehicle license plate
identification system, license plate
location is the first key step. Whether
the plate location is successful or not
and how accurate it is directly
determines the characters segmentation
in the following part. This paper
presents an efficient and smart
approach to the vehicle license plate
location
based
on
mathematical
morphology and variance projection.
The proposed algorithm consists of four
major parts: rough detection of level
location, rough detection of vertical
location, exact detection of license plate
location and extraction of candidate
regions. The outstanding advantage of
this algorithm is simple, smart and fast
as it takes full advantage of open
arithmetic, which can eliminate a large
number of useless details and highlight
the edge information of the license plate.
Keywords-variance projection; iterative
mean filter; open
arithmetic; license plate location
I.
INTRODUCTION
License plate recognition is an
important research topic in the intelligent
transportation systems and has been
extensively used in the systems such as the
electronic police, road surveillance,
vehicle toll, automated car parking
management, and residential intelligence
management systems. The license plate
recognition system generally involves
three key modules, which consist of
license
plate
location,
character
segmentation and character recognition.
How
to exactly and efficiently determine the
license plate

region is a difficult problem in some

conditions such as complex background,
non-uniform illumination, inclined license
plate and the color similarity between the
license plate and background vehicle body.
The position, the plate shape, the
texture feature, the color space feature and
other prior knowledge are the basis of the
plate license location algorithm. A number
of techniques have been proposed for
license plate location, including the
methods based on Histogram Division [1]
[2], comer detection [3][4], vector
quantization [5][6], and color information
[7][8]. But the effect of these location
algorithms always degrades dramatically
when the working environment changes.
This paper proposes an efficient
approach to the vehicle license plate
location
based
on
mathematical
morphology and variance projection. The
experiment proves that this method can
enhance the accuracy of localization.
II.

VARIANCE PROJECTION
FUNCTION
The image projection method has
been proven to be an effective method for
extracting image features an image is
usually represented by two I-dimensiona l
orthogonal projection functions. The
dimension reduction from 2-D to I-D also
reduces the computational load. Owing to
these advantages, the projection method
has been successfully adopted in facial
feature extraction and Chinese character
recognition. Integral Projection is one of
the most popular projection methods.
A. Integral Projection Function
Suppose I (x, y) is the intensity of a
pixel at location (x, y) , the vertical
integral projection v(x, y) and the
horizontal projection H(y) of I(x, y) in
intervals [yl' y2] and [x1'x2] can be
defined as

image

as

in

And their mean vertical and horizontal

projections are defined as
Because eyes are more sensitive to
green, the parameter of green is the
maximum.
As the vertical and horizontal
projection of the license plate present a
continuous distribution of the wave crest
and the wave through, these kinds of
projections functions have been commonly
adopted in some papers.
III.

IMAGE
PROCESSING

PRE-

The pictures of license plates are

generally colourful. If the colourful
pictures are processed directly, it will
lower the processing speed of system.
Therefore, the colourful information of
license plates should be processed to
obtain greyscale pictures. The image preprocessing also includes plate slant
correction to the horizontal line, which is
adopted by the related operation of
mathematical morphology and Radon
transformation.
A. Gray processing
The colourful pictures contained
thousands of bits information. When the
colourful pictures are processed directly,
the execution speed of processing system
will decrease. Therefore, colourful
information of the located plate pictures
should be processed to obtain greyscale
pictures. A pixel of 24-bit BMP images
takes three bytes in memory, which
content red, green and blue color
information. We denote red, green and
blue with R, G and B. The grey value is
showed by g. We have following three
formulas that turn the images into gray

B. Edge
detection
based
on
mathematical morphology
Morphology closing operation can be
used to restrain noises,extract feature,
detection edge and so on. The primary
transformations
of
mathematical
morphology are dilation and erosion.
Opening and closing operation are two
other important morphological operations
composed of dilation and erosion.
Morphology operator
mainly uses
morphology gradient. Though it is
sensitive to the noises, it cant enhance or
magnify noises as in . The most essential
morphology gradient is defined as
following:
The morphology gradient used by
aculeate edge is defined as following
Where A denotes image, B denotes a
structure element. According to the above
description, the morphology erosion edge
detection was presented using the second
morphology operator in the paper.
Furthermore, we used Otsu method to
change grayscale image into binary image.
Then edge detection was performed by
using the morphology erosion algorithm.
We take advantage of the decomposition of
the structuring element object erode the
binary image. The edge detection process
is shown in Fig. 3.1.

C. Horizontal slant vehicle license

plates correction
The
performance
of
character
segmentation is sensitive to skew.
Therefore skew detection and correction
are critical. This problem can be solved
with the Radon transformation. Radon
transformation can be defined in any
dimension variable space region. We
consider the form in two-dimensional
Euclidean space. The transformation is
conducted as below as in :
Where, D denotes the whole image
plane. Eigen function is denoted by Dirac
function. The pixels gray value of a point
(x, y) is noted as f (x, y) in the image. The
distance from origin to straight line is
noted as in x- y domain. is defined as the
angle between the vertical line and x axis,
which is from origin to straight line. The
geometry of the Radon transform is
showed in Fig.3.2 as in . We can see that
the straight line is well-determined when
and are fixed values. In turn, each straight
line in x -y domain represents a point in
domain. So Radon transformation mapped
the straight line in x y domain to a point
in domain.

The Radon transform is closely related

to a common computer vision operation
known as the Hough transform. We can
use the radon function to detect straight
lines of slant vehicle license plates. The
steps are given as following. Step1. Obtain
a binary edge image using the
mathematical
morphology
erosion
algorithm.
Step2. Compute the Radon transform of
the edg e image.
Step3. Find the locations of strong peaks
in the
Radontransform matrix. The
locations of these peaks correspond to the

locations of straight lines in the origin. The

longer the straight line is, the brighter the
corresponding point is. These peak points
are arranged in descending order in Radon
space in view of the veracity. Choose the
front several peaks, calculate the sum of
each row and put them in matrix R .then
run the following orders:
for i = 2 : length(R)
E(i) = a R(i) + (1 a) * E(i 1);
end
Get the maximum of E(i) , a = 90i is
horizonta lrotation angle.
Step4. Lossless rotation correction is
performed to the slant plate in the
horizontal direction. The result of
horizontal slant correction for Fig.3.1 is
shown in Fig.3.3.

IV.
NEURAL NETWORK
A. Artificial Neural Networks
Artificial neural networks are
statistical models of real world systems
which are built by tuning a set of
parameters. These parameters, known as
weights, describe a model which forms a
mapping from a set of given values known
as inputs to an associated set of values, the
outputs. The process of tuning the weights
to the correct values training- is
vehicleried out by passing a set of
examples of input-output pairs through the
model and adjusting the weights in order
to minimize the error between the answer
the network gives and the desired output.
Once the weights have been set, the model
is able to produce answers for input values
which were not included in the training
data. The models do not refer to the
training data after they have been trained;
in this sense they are a functional summary
of the training data [1,10].
The use of artificial neural network
(ANN) in OCR applications can
dramatically simplify the code and

improve quality of recognition while

achieving good performance. Another
benefit of using neural network in OCR is
extensibility of the system i.e. ability to
recognize more character sets than initially
defined. In the character recognition
algorithm using neural networks, the
weights of the neural network were
adjusted by training it using back
propagation algorithm. The size of each
character is 28-by-18 pixels which are
arranged column wise to give 504 _ 1
arrays as input. In order to train the neural
network, we have created different sets
each containing digits from 0 to 9. This is
called Block training. The complete net
work was implemented as a library, which
was statically tied to the project. This
helped to isolate the neural network code
from the rest of the preprocessing and
segmentation code . It also helped to
reduce the memory required for the
program.Character recognition of the
number-plate is a fairly well developed
field in computer vision in which template
matching and neural networks are often
used and can produce satisfactory results
However, template matching has its
drawbacks in some aspects comparing
with neural networks. For example, when
characters of number-plate are segmented,
neural network approach is preferred to
template match one due to more
computation cost of the later method .
As a very powerful technique for
pattern recognition problems, it is not
surprising that neural
networks became the basis of a number of
such systems
B. The Back Propagation (Bp)
Algorithm
The back propagation is a widely
used algorithm, and it can map non-linear
processes. It is a feed forward network
with the one or more hidden layers. The
elementary architecture of the back
propagation network has three layers.
There are no constraints about the number
of hidden layers. Back propagation is a
systematic method for training multilayer

artificial neural net works. It has a

mathematical foundation that is strong if
not highly practical.
Using
neural
network
has
advantage from existing correlation and
statistics template techniques that allow
being stable to noises and some position
modifications of characters on license
plate. The network must memorize all the
Training Data (36 characters). For the
validation of the network we have built a
program that reads the sequence of
characters, to cut each character and resize
it and put the result.
V.

CONCLUSION

1. A database including 100 images

taken by CCD camera under
different conditions and 126
images Downloaded from the
Internet was used in the
experiment. Only in three images,
the license plate was not able to be
located. The positioning accuracy
is 98.6%. One reason is that the
estimating value of L mentioned in
open arithmetic is not precise.
Another reason lies in that the
undetected license plates were too
unclear and lacked effective texture
information
2. Because of character recognitions
are generally very sensitive to skew
plate. Therefore, skew detection
affect recognition rate of license
plate. If the images of license plate
do not have skew, the recognition
rate will greatly be improved. In
addition, the recognition rate of
English letters and numeral are
high then that of Chinese letter.
The result of the experiment, we
can see the proposed approach is
robust. Yet there are still some
images failed to detect in the
experiment
(especially
skew
images). We will improve the
performance of the algorithm
through skew correction of license

plate further and Chinese letter

recognition of license plate.
3. This paper introduces a neural
network approach for automatic
identification of vehicle License
plate. Recognition of character are
better then Template matching
method. If the accuracy of the
results is a critical factor for a
vehicle number plate recognition
application, then the network
having many hidden layers should
be used but if training time is a
critical factor then the network
International
Journal
on
Cybernetics & Informatics ( IJCI)
Vol.2, No.1, February 2013 having
single hidden layer should be
used.The proposed approach of
license plate recognition can be
implemented by police to detect
speed violators, parking areas,
highways, bridges or tunnels. Also
the prototype of the system is
going to be integrated and tested as
part of the sensor network.
Currently
there
are
certain
restrictions on parameters like
speed of the vehicle, script on the
number plate, cleanliness of
number plate, quality of captured
image, skew in the image which
can be removed by enhancing the
algorithms further.
REFRENCES
1. Wang, G. Gao, and H. Yang,
"Research and implementation of
license plate location based on
histogram division method," The
Ninth International Conference on
Electronic
Measurement
&
Instruments, pp. 231-233, 2009.
2. F. Yang, and Z. Ma, "Vehicle
license plate location based on
histogramming and mathematical
morphology," The Fourth IEEE
Workshop
n
Automatic

Idenfication
Advanced
Technologies, pp. 89-94,2005.
3. H. Huang, G. Ma, and Y. Zhuang,
"Vehicle license plate location
based on Harris comer detection,"
IEEE Inte
4. Z. Qin, S. Shi, and 1. Xu, "Method
of license plate location based on
comer feature," The Sixth World
Congress on Intelligent Control
and Automation, vol. 2, pp. 86458649, 2006.
5. R. Zunino, and S. Rovetta, "Visual
location of license plates by vector
quantization," Proceedings of the
1999
IEEE
International
Symposium on Circuits and
Systems, vol. 4, pp. 135-138,1999.
6. 6. R. Zunino, and S. Rovetta,
"Vector quantization for license
late location and image coding,"
IEEE Transactions on Industrial
Electronics, vol. 47, pp. 159167,2000.
7. 7. M. M. Maged Fahmy,
Automatic
number-plate
recognition:
neural
network
approach, Proceedings of Vehicle
Navigation
and
Information
Systems Conference, Sept. 1994,
pp.99-101,
8. T. Kato, Y. Ninomiya, I. Masaki,
Preceding vehicle recognition
based on learning from sample
images, IEEE Transactions on
Intelligent Transportation Systems
vol. 3 (4), 2002, pp. 252260.
9. Otto Fucik, Pavel Zemcik, Pavel
Tupec, Ludek Crha, Adam Herout,
"The Networked hoto-Enforcement
and Traffic Monitoring System of
rollout steps includes: Unicam,"
Proceedings of the 1th IEEE
International Conference on the
Engineering of Computer-Based
Systems, 2004, pp. 234251..
10. Rafael and Richard,2009, Digital
Image Processing, 3rd ed.
Prentice-Hall Inc.

11. Hui Wu and Bing Li,2011,

License
Plate
Recognition
system, IEEE.
12. K.
Kanayama,et
al.,1991,
Development of vehicle-license
number recognition system using
realtime image processing and its
application
to
travel-time
measurement, Processings of
IEEE
Vehicular
Technology
Conference, pp.789-804.
13. D.U. Cho and Y.H. Cho,
Implementation of pre processing
independent of environment and
recognition of car number plate
using histogram and template
matching, The Journal of the

Korean Institute of Communication

Sciences, 23.
14. Parul Shah, et al.,2009, OCRbased Chassis-Number Recognition
using Artificial Neural Networks,
ICVES 2009.
15. Feng Yang, and Fan Yang,
Character Recognition using
Parallel BP Neural Network,
IEEE.
16. A. Akoum, et al.,2009, Two Neural
Networks for License Number
Plates Recognition, Journal of
Theoretical
and
Applied
Information Technology.

TUGAS UTS PENGANTAR SKRIPSI

MEREVIEW 5 JURNAL

I GUSTI NGURAH AGUNG DWI JAYA PUTRA

NIM 1304405054

JURUSAN TEKNIK ELEKTRO

FAKULTAS TEKNIK UNIVERSITAS UDAYANA
BUKIT JIMBARAN
2016