Sie sind auf Seite 1von 19

INTRODUCTION

One of the most significant factors which contribute to road


accidents is drowsiness while driving. According to the
NATIONAL HIGHWAY TRAFFIC SAFETY
ADMINISTRATION[NHTSA] ,drowsy driving accounts for
around 100,000 crashes, resulting in 1,550 deaths and
40,000 injuries annually. Drowsiness can be defined as a
nearly sleepy state caused by tiredness. This state is
regarded to be an undesired one while driving owing to the
fact that it reduces mental alertness and slows down the
reaction time to critical conditions on road , therby
impairing the ability of the driver to operate the vehicle
safely and enhancing the risk of road accidents. It is
therefore , from both cost and safety point of view,
necessary as well as beneficial to develop a SYSTEM FOR
DETECTION OF DOZING DRIVERS.

ABSTRACT

The basic aim of our project is to design a real time


system for detecting drowsiness at the wheel. Initially
one or two cameras are employed to snap a few pictures of
the drivers face at certain time intervals. This is followed
by what we term as Image Processing ,to accomplish
which we are resorting to the MATLAB-a high level
computing language, with an Image Processing Toolbox
that assists us in data analysis and interpretation. Image
Processing serves the purpose of detecting visual features
like blinking of the eye which are indicative of the dozing
state of the driver. The system terminates with the
generation of an alarm in the form of light or sound as soon
as the dozing state is detected.

BASIC BLOCK DIAGRAM OF THE


PROPOSED SYSTEM

TECHNIQUES ADOPTED IN
DROWSINESS DETECTION
Owing to the fatalities caused by drowsiness on wheel,
research activities are being conducted all over the world
to combat this problem and this has led to the evolution of
several drowsiness detection mechanisms. Some of them are:-

PREVIOUS WORKS UNDERTAKEN

NAME OF THE
YEAR OF
RESEARCHER(S PUBLICATION
)

DETECTION METHOD
PROPOSED

Hyun J.B., Gih S.C., Ko


K.K., Kwang S.P.

November 2011

A smart health monitoring chair


for nonintrusive measurement
of biological signals using
EEG,ECG etc.

Forsman P.M., Vila B.J.,


Short R.A., Mott C.G.,
van Dongen H.P.A.

May 28,2012

Abe T., Nonomura T.,


Komada Y., Asaoka S.,
Sasai T., Ueno A.,
Inoue Y.
Taranpreet Kaur,
Sanjay Kumar Singh

October 2011

Nawal Alioua, Aouatif


Amine, and Mohamme
d Rziza

July 2014

Vehicle based measurements


by use of sensors in vehicle
components like steering wheel
and analyzing the signals
received from the sensors
Evaluation of percentage of
closure time (PERCLOS) of the
eyelid with decrease in eye
blink frequency.
During drowsiness the eyes
gets swelled. This drowsiness
detection system is used to
detect the swelling of soft
tissues.
Drivers fatigue detection
based on yawning extraction
that is based on mouth
detection approach.

March 2014

Ayush Joshi , Shruti


Gujrati , Amit Bhati

Qiang Ji and Xiaojie


Yang

2002

Head position detection


technology which simply
determines the head tilt angle
which can detect the
drowsiness of a driver.
A detection drowsiness system
based on infrared light
illumination and stereo vision.
This system localizes the eye
position using image

differences based on the bright


pupil effect.

Commercial Technologies
1.

2.

AAlert - AAlert, a driver fatigue wrist device has been developed by Dan Ruffle. This uses an
accelerometer to detect motion from your wrist. Whilst on a drive and your wrist is not moving a
vibration is sent to your wrist and AAlert requires a movement to then stop the vibration. AAlert
then records your reaction times over a period of time to determine whether or not you should be
on the road.

Audi Rest recommendation system

3.

BMW:Active Driving Assistant with Attention Assistant analyses driving


behaviour and, if necessary, advises the driver to rest. The advice to take a break is
provided in the form of graphic symbols shown on the Control Display.

4.

Bosch: "Driver drowsiness detection takes input from the steering angle sensor,
front-mounted lane assist camera, vehicle speed and turn signal stalk.

5.

Ford:Driver Alert

6.

Mazda: Lane Departure Warning System

7.

Mercedes-Benz: Attention Assist In 2009, Mercedes-Benz unveiled a system called


Attention Assist which monitors the driver's fatigue level and drowsiness based on his/her
driving inputs. It issues a visual and audible alarm to alert the driver if he or she is too
drowsy to continue driving

8.

Nissan: Driver Attention Alert (DAA)

9.

Subaru: EyeSight Driver Assist

10.

Volkswagen: Fatigue detection system

11.

Volvo Cars: Driver Alert Control In 2007, Volvo Cars launched the world's first
Driver Drowsiness Detection system, Driver Alert Control. The system monitors the car's
movements and assesses whether the vehicle is being driven in a controlled or
uncontrolled way.

12.

Anti Sleep Pilot - Danish device that can be fitted to any vehicle, uses a
combination of accelerometers and reaction tests.

13.

Vigo - Smart Bluetooth headset that detects signs of drowsiness through the eyes
and head motion, and uses a combination of light, sound and vibration to alert the user.

COMPARATIVE STUDY

Method Used:
ECG and EEG

Steering Wheel
Movement

Advantages
1.

Accuracy is very high.

2.

Response is very fast.

1.

Designing such a system


is very easy.

2.

Works very well in


very limited
situations. They
can function
reliably only at

Disadvantages
Sensing electrodes
would have to be
attached directly
onto the drivers
body, and hence
be annoying and
distracting to the
driver.
2. If it is to be
developed as a
non-invasive
method, cost will
be too high.
1.

1.

Too dependent on
the geometric
characteristics of
the road and to a
lesser extent on
the kinetic
characteristics of

Computer vision

particular
environments.
3. Non-invasive.
4. Response speed is
slow.

the vehicle.
2. Accuracy and
speed of response
is not very good.

1.

Detection speed is fast.

1.

2.

Non-intrusive method.

robustness in the face of

3.

No extra installation of

changing environmental

hardware required.

conditions, such as

4.

Image processing
algorithms does not require

shadows and lighting.

2.

Noise might corrupt the


information received.

upgrading tags, identifiers,


or transponders devices on

Difficult to maintain

3.

Moderate accuracy.

vehicles.

5.

Economical and precise.

BASIC PROCESS INVOLVED IN


DETECTION PROCESS
The fundamental concept we have tried to apply in our
method of drowsiness detection is recognizing the
elliptical shape of human eye.
Drowsiness results in a partially closed state of the eye.
This means that the white portion of the eye known as the
sclera shall not be fully visible or rather it will be absent in

the different images of the eye snapped at a certain time


interval by the webcamera.

In this regard it is to be noted that that drowsiness shall


be concluded only if the sclera is not visible in at least 3-4
images .This is because absence of sclera in one single
image might be indicative of the blinking of the eye as a
regular body activity.
As soon as the system detects the drowsy state, it will
trigger an alarm which will alert the driver of the fact that
he is not in a fit state to drive.

PSEUDOCOE FOR SIMPLIFIED


UNDERSTANDING OF THE
PROCESS
1.
Specify the resolution of the image to
be analyzed.
2.
Read the image from the file where it
has been stored.
3.
Display the original RGB image.
4.
Convert the RGB image to a gray
image.
5.
Display the gray image.
6.
Threshold the gray image.
7.
Convert the gray image to a binary
one.
8.
Invert the binary image.
9.
Display the binary image.
10. Trace the outer boundaries of the image.
11. Determine the properties of the object under
analysis.(eccentricity)
12. If the value of the eccentricity lies between
0 and 1,an ellipse is detected.
13.

OUTPUTS OBTAINED AS PER THE


PRELIMINARY FRAMEWORK OF THE
PROGRAM

PRELIMINARY FRAMEWORK OF THE


PROGRAM

AREAS YET TO BE EXPLORED


Our focus in this project is on devising a method to ensure
that the process of ellipse detection (detecting the eye)
becomes faster and more simplified. We aim at exploring the
property of eccentricity of an ellipse to ensure accurate
detection of the eye and then look for whether the eye is
open or not. If the entire process of ellipse detection can be
made accurate enough only by virtue of measuring the
eccentricity, it shall reduce the memory and time required
for drowsiness detection. Till now , we take into
consideration the very fundamental concept of any object
having an eccentricity between 0 and 1 to be regarded as an
ellipse, which always might not be true. Further research
shall aim at improving upon this concept and trying to evolve
a better option.
We are yet to look into the various ways of interfacing an
efficient alarm system in the form of indicators/sound system
or combining both to give a final shape to the entire process
of drowsiness detection by timely intimation to the driver
about his unfit state

COST PREDICTION

REFERENCES
1. Hyun J.B., Gih S.C., Ko K.K., Kwang S.P. A Smart health monitoring chair for
nonintrusive
measurement of biological signals. IEEE Trans. Inform. Technol.
Biomed. 2012;16:150158
2. International Journal of Engineering Research & Technology (IJERT) vVol. 2 Issue
1, January- 2013 ISSN: 2278-0181

ACKONWLEDGEMENT

We desire to express our acknowledgement towards the


department of APPLIED ELECTRONICS AND
INSTRUMENTATION ENGINEERING of HERITAGE
INSTITUTE OF TECHNOLOGY,KOLKATA for encouraging us
to progress with our initiative. In this regard,we shall also
thank our project mentor, Prof. Surajit Bagchi whose valuable
advice and guidance is assisting us to give our project a
desired shape.

CONTENT
Introduction
Abstract
Basic Block Diagram Of The Proposed System
Components To Be Used
Technique Adopted In Drowsiness Detection
Previous Works Undertaken
Commercial Technologies
Comparative Study
Basic Process Involved In Detection Process
Pseudocode For Simplified Understanding Of The Process
Preliminary Framework Of The Program
Output Obtained From Preliminary Framework Of The Program
Areas Yet To Be Explored
Cost Prediction
References

COMPONENTS TO BE USED

HARDWARE
One or more web cameras
For a clearer and vivid image, a webcam with more
megapixels should be used. It is desirable to use a
webcam with around 640X480 pixels.
Frame Rate determines the number of images per
second which can be displayed. For a standard
webcam, frame rate of 30fps (frame per second)
is adequate.
A glass lens is preferred over plastic lens.
It should be enabled with continuous autofocus so
that it is not diverted towards other objects within
the frame.
The webcam should be able to maintain a good
image quality inspite of poor light conditions.
Laptop(or on-board computer)
This provides the platform for the actual
processing task to be accomplished by the
requisite software.
Alarm(light and/or sound)

Contd....

SOFTWARE
Matlab (Matrix Laboratory)
Java

Das könnte Ihnen auch gefallen