IJISET - International Journal of Innovative Science, Engineering & Technology, Vol. 2 Issue 5, May 2015.
www.ijiset.com
ISSN 2348 – 7968
An Innovative System for Accident Detection, Warning
and Prevention
Geetanjali Shintre[1], Gowrima E.[2], Prabhudev R.[3], Shilpa M.S.[4]
Mrs. Bhat Geetalaxmi Jayram (Associate Professor)[5]
Department of Information Science and Engineering
The National Institute of Engineering, Mysore, Karnataka, India
ABSTRACT
This application will be installed on
the smartphone. Every person using a
smartphone has the GPS facility inbuilt which
is in turn connected to the vehicle by which a
person can find his location. In this application
we can predict traffic by looking at the traffic
of GPS at a particular place. Since now we can
predict the traffic, it can be avoided by taking
another route to the destination. The systems
should be able to identify each vehicle and
track its behaviour, and to recognize situations
or events that are likely to result from a chain
of such behaviour.
Using image processing we detect for
accidents. A camera will be installed in the
vehicle which keeps track of the face
expressions. If there is sudden change in the
expression of the user, it detects there is an
accident. The system will then send the
accident location acquired from the GPS along
with the time by utilizing the GSM network.
This will help to reach the rescue service in
time and save the valuable human life.
Keywords
— Accident Detection, Image
processing, Camera, shock analyses
1. INTRODUCTION
Traffic Prediction and Accident
Detection using GPS and GSM
Technology” is an android application
which is to be installed over a smartphone
and the application is very handy since
now-a-days almost everybody owns their
own smartphone.
Traffic prediction refers to
avoiding the traffic from any remote
location or keeping yourself updated about
any traffic jams on the route passing
through. Using the GPS initially, location
of the particular vehicle is found which
checks for any traffic jam on the way.
Since every vehicle’s location is detected
by the GPS, with the help of GSM
technology the application will inform to
user to take another route and hence can
save their precious time.
Accident detection is another
module of this project. Using the image
processing technology the accident is
detected. A camera is to be fixed in every
vehicle which captures continuous images
of the face expression of the driver.
Overlapping of images is considered, when
there is no much movement or if the
images are to be somewhat similar then it
is said to be that everything is going well.
Suppose if there is sudden drastic change
in the face expression of the driver the
accident is then detected. If so happens,
then there will be a message displayed by
the application which will accept one input
to the question asked to the driver if he’s
“OK”. If the input is given “Yes” since the
accident is minor, we know that the driver
272 IJISET - International Journal of Innovative Science, Engineering & Technology, Vol. 2 Issue 5, May 2015.
www.ijiset.com
ISSN 2348 – 7968
has no much harm. A timer will be set to
give this input. Suppose there is a major
accident, the user won’t be able to give
any input, the timer triggers; automatically
the input will be then taken as “NO” and
immediately the police control room or the
ambulance will be informed as well.
2. PROPOSED SYSTEM
We reviewed driver face monitoring
systems from different viewpoint in the
previous sections. In this section, a coarse
comparison of some of the reviewed
systems is presented from the discussed
viewpoints. Also, we explain the critical
problems in the current systems.
However
the
method
for
fatigue/distraction detection is effective to
achieve a robust system, but current driver
face monitoring systems suffer from two
main problems:
Robust and precise detection and tracking
of face and facial components, Robust and
precise symptom extraction. These
challenges are:
Developing
illumination
invariant
algorithms for detection and tracking of
face and facial components
Developing algorithms for detection and
tracking of face and facial components
with robustness against skin color and
partial occlusion (e.g., glasses, sun-glasses
and whisker)
Developing robust algorithms for driver
face tracking with respect to head rotations
in different directions
Developing algorithms for symptom
extraction with robustness against different
races (e.g., eyelid distance of East Asian
people is low on normal state)
Developing algorithms for symptom
extraction from eye region with robustness
against wearing glasses and sun-glasses.
Developing color invariant algorithms for
symptom extraction from mouth region
with robustness against skin-color and
occlusion (e.g., whisker)
Fast processing to achieve real-time
systems
Tracking of the eyes: We track the eye by
looking for the darkest pixel in the
predicted region. In order to recover from
tracking errors, we make sure that none of
the geometrical constraints are violated. If
they are, we relocalize the eyes in the next
frame. To find the best match for the eye
template, we initially center it at the
darkest pixel, and then perform a gradient
descent in order to find a local minimum
Figure 1: Flow of data through the system
273 IJISET - International Journal of Innovative Science, Engineering & Technology, Vol. 2 Issue 5, May 2015.
www.ijiset.com
ISSN 2348 – 7968
REFERENCES
Figure 2: Steps of image processing
[1]
N. Peterfreund, “Robust tracking of
position and velocity with kalman
snakes,” IEEE Trans. Pattern Anal.
Machine Intell. , vol. 21, pp. 564–
569, June 1999.
[2]
M. Kass, A. Witkin, and D.
Terzopoulos,
“Snakes:
Active
contour models,” Int. J.Computer
Vision, vol. 1, pp. 321–331, 1988.
[3]
R.-T. Non, - Intrusive monitoring
and prediction of driver fatigue by
Q. Ji, Z. Zhu, and P. Lan, IEEE
Transactions
on
Vehicular
Technology, vol. 53, no. 4, July
2004.
[4]
J. Healey and R. Picard, SmartCar: Detecting driver stress, in
Proc. of 15th Int. Conf. Pattern
Recognition, Barcelona, Spain,
2000, vol. 4, pp. 218–221.
274