Lane Departure Warning System
Uploaded by
Nayan GadreLane Departure Warning System
Uploaded by
Nayan GadreDevelopment of
Lane Marker Detection & Warning
system.
Project by:- Gadre Nayan A (B3073016)
Awachat Prasanna K(B3073002)
External Guide Internal Guide
Prof. H.K.Sawant Prof. Mrs. M.P.Atre
HKS systems. PVG’s COET.
Near Karishma apts.
Kothrud. Pune.
OBJECTIVE
To design a system which can detect the lanes on the road from
the images taken through a camera mounted on the moving
vehicle and give a warning signal corresponding to its safety level
with respect to the current position.
OVERVIEW
I. Introduction
II. Literature Survey
III. Specifications
Hardware
Software
IV. Block Diagram
V. Present status
VI. References
WHY?
Financial
Damage Expected Rise :
15% of World 60% by 2020
GDP
Road Accidents Lane
In INDIA approx 270
Departure
People die each day
LANE
DEPARTURE
WARNING
SYSTEM 40 % of all
fatalities
INTRODUCTION
1. Types: Rear end collisions,
Side swipes, Head on
collisions, Accidents while
Merging or Changing Lanes,
Driving off the road.
2. Most occurrences result from
Distraction , Inattention and
Driving Fatigue of driver.
3. A lane departure warning is a
safety system in a car which
alerts the driver when the car
drifts out of its lane.
SPECIFICATIONS
• Image Acquisition using
Camera (320x240). • 8 bit uc for control of alarm,
Intex 300k CMOS sensor. Led indication and motors.
• Indication using Buzzer.
• Image processing and
• USB interface .
control of hardware using
• Rechargeable SMF Battery VB 6.0/MATLAB.
(12V) powered systems.
• Driving 9V DC 2kg-cm motors.
(to demonstrate tracking).
BLOCK DIAGRAM
LANE
Power
Supply
COMPUTER uC
Sensor selection
CMOS CCD
• Lesser response time. • More. Shifting of charge from
• Cheaper each pixel takes time.
• CMOS imagers offer superior • CCDs offer slightly better image
integration, quality and flexibility at the
• Most suitable for medium quality expense of system size.
video applications • They remain the most suitable
• Avoids Vertical smear. technology for high-end imaging
applications such as space
photography.
• Anti blooming arrangement
needed.
Color Models
• An 24bit-RGB image, also referred to as a "true colour"
image.
• HSI (Hue saturation intensity) format suited for machine
vision applications(robotics) gives perceptually equivalent
image.
• CYMK (cyan yellow magenta black) format used in printers,
coloured Xerox etc.
LANE DETECTION BY IMAGE PROCESSING
Steps involved in lane detection are
•Conversion of RGB image to an HSI image
•Thresholding
•Image Filtering
•Detection of appropriate lines corresponding to the lane
FUTURE WORK & LIMITATIONS
The system may not work under the conditions mentioned below
•Very bad weather conditions like heavy fog which reduce visibility.
•Yet to decide the practicality of this system in Automatic tracking.
Although demonstration for tracking as a mock is possible.
FUTURE PROSPECTS
• This concept can be implemented as a subsystem of a futuristic
fully automated vehicle.
• It can be linked to the Electronic(Engine)Control Unit of the
vehicle which will take the output of the system and perform the
control action
Completion Status
• Serial communication program tested on
AT89C51.
• Image Display, pixel data extraction, grayscale
conversion algorithms complete in VB 6.0.
• A test program for RGB to HSI conversion and
grayscale conversion tested in Matlab.
REFERENCES
• Visual Basic 6.0 The Black Book. By Steven Holzner.
• Visual basic game programming:- by Jonathan harbour.
• Image Processing Analysis and Machine Vision :- by MILAN
SONKA , ROGER BOYLE.
• Kristijan Macek , Brian Williams :- “ A lane detection vision
module for Driver assistance”
• Image processing using Matlab :- Gonzalez, Woods, eddins.
• www.ftdichips.com
• www.atmel.com
• www.msdn.com
• www.devx.com
• forums like:-
edaboard,xtremeVBtalk.
THANK YOU
RGB IMAGE
• An RGB image, also referred to as a "true colour" image.
• It is stored as an RxGxBx8bit data array that defines red,
green, and blue colour components for each individual pixel.
• The colour of each pixel is determined by the combination of
the red, green, and blue intensities stored in each colour
plane at the pixel's location.
GRAYSCALE IMAGE
• An intensity image (or grayscale image) is a two dimensional data
matrix.
• The values of the matrix represent intensities within the range 0
to 255.
• Each element of the matrix represents an image pixel. The
intensity 0 represents black and the intensity 255 represents
complete white.
• There are many ways to convert an RGB image to grayscale.
Averaging:- G= (R+G+B)/3.
Luminance:- G= 0.3xR + 0.6xG +0.1xB.
RGB TO GRAYSCALE IMAGE
RGB images can be
converted to grayscale
by eliminating the hue
and saturation
information while
retaining the luminance.
‘rgb2gray’ command in
MATLAB converts RGB
values to grayscale
values by forming a
weighted sum of the R,
G, and B components. A Gray scale Image
We use Y= 0.3R+0.6G+0.1B.
THRESHOLDING
• Thresholding is carried out on the grayscale image to enhance its
contrast.
• A certain threshold level is decided and pixels having intensity
above that level are treated as white(255) while pixels below the
level are treated as black(0).
• This method will help in extracting the lane data easily.
• However, not only lane markers but certain surrounding objects
whose intensity levels lie above the threshold, may also get
emphasised. Hence the next process is applied.
A Grayscale Image after Thresholding
IMAGE FILTERING
• Filtering is a technique for modifying or enhancing an
image. It is a neighbourhood operation, in which the value
of any given pixel in the output image is determined by
applying some algorithm to the values of the pixels in the
neighbourhood of the corresponding input pixel. This has
the effect of smoothing out local highlights and blur edges
in an image.
EDGE DETECTION
• The basic idea behind edge detection is to find places in an image
where the intensity changes rapidly.
• Edge detection is process of generating a binary image from a
grayscale image with pixel values as 1 where there are edges and 0
elsewhere.
• Weaker edges in the image are ignored by specifying a threshold of
gradient.
