Reference 120315c08

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/313106893

DYNAMIC TRAFFIC SIGN BOARD DETECTION AND RECOGNITION USING SVM


CLASSIFIER AND EXHAUSTIVE TEMPLATE MATCHING - AN INTERACTIVE
TEXT/SYMBOL TO VOICE CONVERTER

Article · March 2015

CITATIONS READS

0 277

5 authors, including:

Anand Kumar M
Indian Institute of Science
15 PUBLICATIONS   23 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

QoS-Aware Routing in Wireless Body Sensor Networks View project

Design of Energy-efficient Optimal Framework in Short Computation Time for Dynamic Wireless Body Area Network Problem View project

All content following this page was uploaded by Anand Kumar M on 31 January 2017.

The user has requested enhancement of the downloaded file.


INTERNATIONAL JOURNAL OF ADVANCED RESEARCH IN COMPUTER SCIENCE AND APPLICATIONS
ISSN 2321-872X ONLINE ISSN 2321-8932 PRINT IMPACT FACTOR : 0.782
VOLUME 3, ISSUE 3, MARCH 2015.

DYNAMIC TRAFFIC SIGN BOARD DETECTION AND


RECOGNITION USING SVM CLASSIFIER AND EXHAUSTIVE
TEMPLATE MATCHING - AN INTERACTIVE TEXT/SYMBOL TO
VOICE CONVERTER
ARPITHA B.R1 DIVYASHRE S2 SHEEMA S3

VICKNESH K4 ANAND KUMAR M5


1, 2, 3, 4 & 5
Department of Computer Science Engineering, NIE Institute of Technology, Mysore

ABSTRACT
Traffic Sign Detection and Recognition is a driver support function which can be used to notify and
warn the driver which restrictions may be effective on the current stretch of road. Examples for such
regulations are „speed limit zones‟ or „no-overtaking‟ indications.In this paper dynamic approach for

r g
detecting and recognizing sign boards is presented. Genetic algorithm (Exhaustive template matching)
is used for detection and recognizing of road signs. SVM classifier is used to classify recognized road
signs.

. o
OCR algorithm.
s a
Keywords: Sign board detection, Exhaustive template matching, SVM classifier, Genetic algorithm,

Introduction
r c of traffic signs is, therefore, important for

i. j
There are certain rules that have been prepared
for the benefit of people and the idea of
preparing these rules is not that they should be
a automated driving or driver assistance systems.

Literature survey

w
understood by the drivers, but it should also be
understood by the cyclists, pedestrians and other
Genetic algorithms are based on natural
selection discovered by Charles Darwin [1].

w
people. It is essential to follow all the rules and They employ natural selection of fittest

w
regulation and they are clearly listed here.
People are recommended that they should be
carefully observing all the rules and regulation
individuals as optimization problem solver.
Optimization is performed through natural
exchange of genetic material between parents.
and it is effectual to be careful, considerate and Offsprings are formed from parent genes.
patient. These signs have got the very prominent Fitness of offsprings is evaluated. The fittest
role to play in the traffic system and they are individuals are allowed to breed only.
made for the safety of people.
In machine learning, support vector
Road signs define a visual language that machinesare supervised learning models with
can be interpreted by drivers. They represent the associatedlearning algorithms that analyze data
current traffic situation on the road, show danger and recognize patterns, used for classification
and difficulties around the drivers, give them and regression analysis. Color-based methods
warnings, and help them with their navigation search for inconsistencies in theinteractions
by providing useful information that makes between object color and light color [2][3].
driving safe and convenient. Traffic signs The road sign recognition system is not the
provide the driver various information for safe only subsystem in vision-based driver assistance
and efficient navigation. Automatic recognition systems. There are several other subsystems,

www.ijarcsa.org 31 admin@ijarcsa.org
INTERNATIONAL JOURNAL OF ADVANCED RESEARCH IN COMPUTER SCIENCE AND APPLICATIONS
ISSN 2321-872X ONLINE ISSN 2321-8932 PRINT IMPACT FACTOR : 0.782
VOLUME 3, ISSUE 3, MARCH 2015.

Which perform obstacle recognition, and the relationships among subsystems and their
environmental change detection, etc. Developing contribution to meeting a common goal.It
and integrating these subsystems to collect provides a framework for visualizing the
significant information in driving environments organizational and environmental factors that
is very important for improving traffic safety. operate on a system.
They hoped that the proposed DVM will be
helpful for designing various subsystems for Functional Requirements
functions [4]. A. Pre-processing Block
Pre-processing will load the image as well as
System Analysis
check contrast, brightness, and clarity. The
System analysis and design relates to shaping following figure shows flow of control in the
organizations, improving performance and pre-processing block.
achieving objectives for profitability and
growth. The emphasis is on systems in action,

r g
. o
sa
Fig. 1 Pre-processing block

r
If these parameters are off from our desired cB. Image Processing and Recognition

i. j a
values for these, adjustments will be performed.
If the software cannot obtain the contrast or
brightness needed, it may not be able to identify
Pre-processing shall check contrast,
brightness, and clarity. This block shall make
sure the image is ready to have image processing

w
if there is a sign in the image. done to it. After passing through this pre-
processing block, the image shall be ready to

w have processing algorithms applied to it.

w
Fig. 2 Steps for Processing and Recognition
The application of processing algorithms shall allows for the system to identify exactly what
take the pre-processed image and find colours of sign is contained in the image that was
interest and look for shapes relating to the sign processed.
or signs we are searching for. This block shall
Non Functional Requirements:
find regions of interest on the image and these
 Thesoftware to be developed must detect
shall be further processed to obtain the type of
only road sign boards.
sign. This is done in the following block.
 The software to be developed must ignore
The classify sign block shall take the
all other objects except road sign boards.
regions of interest passed from the algorithms
 The software to be developed must
block. These regions shall be analysed and used
recognise the road signs correctly.
to compare to „templates‟ of known signs. This

www.ijarcsa.org 32 admin@ijarcsa.org
INTERNATIONAL JOURNAL OF ADVANCED RESEARCH IN COMPUTER SCIENCE AND APPLICATIONS
ISSN 2321-872X ONLINE ISSN 2321-8932 PRINT IMPACT FACTOR : 0.782
VOLUME 3, ISSUE 3, MARCH 2015.

 The software to be developed must display System Design


the road sign in textual format. Camera-based road sign recognition (RSR) is a
 The software to be developed must convert valuable tool for any drivers, particularly with
the text output to voice output. respect to signs indicating danger and speed
limits. It is able to recognise selected road signs,
identify their meaning and assist the driver in
adapting his or her driving style to the situation
and regulations.

r g
. o
sa
r c
i. j a
w
w
w Fig. 3 System design methodology

Feature Based Detection: signs. This list is forwarded to the recognizer for
further evaluation, and then to the classifier to
In computer vision and image
decide whether the objects in the list are either
processing the concept of feature detection
rejected objects or road signs, and in this case
refers to methods that aim at computing
the classifier responds with a sign code. The
abstractions of image information and making
classifier will make use of machine learning
local decisions at every image point whether
technique called support vector machine (SVM)
there is an image feature of a given type at that
to identify the candidate.
point or not. The resulting features will be
subsets of the image domain, often in the form
Implementation and results
of isolated points, continuous curves or
connected regions. In this experimental setup, a camera is mounted
inside a car. The video captured by the camera
Recognition and Classification:
consists of a number of frames depending on the
The output of the detection stage is a list of camera. For each frame a thread is created to
candidate objects that could be probable road process it simultaneously. By which the system

www.ijarcsa.org 33 admin@ijarcsa.org
INTERNATIONAL JOURNAL OF ADVANCED RESEARCH IN COMPUTER SCIENCE AND APPLICATIONS
ISSN 2321-872X ONLINE ISSN 2321-8932 PRINT IMPACT FACTOR : 0.782
VOLUME 3, ISSUE 3, MARCH 2015.

detects whether a road sign is present or not converted to voice. Our major motivation in this
using genetic algorithm (exhaustive template project is to alert and give the road sign
matching). Any text or number present is information to the driver in time which saves
detected by optical character recognition life.
algorithm. Detected road sign is classified using
SVM classifier. The language used is c#.NET References
framework 4.0 [5]. [1] H.P Schwefel, G. Rudolph. Contemporary
The system after recognizing the traffic signs it evolution strategies.Advances in artificial life,
displays the traffic sign in a picture box and a 1995, 893 – 907.
text output is show on screen .The text is then [2] C. Riess and E. Angelopoulou, “Scene
converted to voice output. illumination as an indicator of image
Applications and conclusion manipulation,” Inf. Hiding , vol. 6387, pp. 66–
80, 2010.
The system acts as an advanced warning system,
intimating the driver on the traffic sign boards [3] S. Gholap and P. K. Bora, “Illuminant colour
present on the current stretch of road. This
system is one of the basic modules necessary for
r g
based image forensics,” in Proc. IEEE Region
10 Conf., 2008, pp. 1–5.
the next generation of cars. These next
. o
[4] "A ROAD SIGN RECOGNITION SYSTEM

a
generation cars are automated and require less
BASED ON DYNAMIC VISUAL MODEL", C.
interaction with the driver.

It is concluded as the road signs can be


c s
Fang, C. Fuh, S. Chen, and P. Yen.

detected and recognized using the system. The

a r [5]http://www.tutorialspoint.com/csharp

i. j
output of the recognized sign board will be

w
w
w

www.ijarcsa.org 34 admin@ijarcsa.org

View publication stats

You might also like