You are on page 1of 15

ROAD LANE DETECTION

OF AUTONOMOUS CAR

ANURAG GAUTAM (211210013)


DEVANSH SRIVASTAVA (211210022)
KULBIR SINGH (211210034)

Under the guidance of Dr. CHANDRA PRAKASH SAHU


PROBLEM STATEMENT

AS WE KNOW THAT IN FUTURE AI WILL PLAY A VITAL ROLE IN THE WORLD AND ONE
OF
ITS APPLICATION WOULD BE SELF DRIVING CARS I.E AUTONOMOUS CARS SO TO
PREVENT THEM FROM ACCIDENTS OR CAUSING DESTRUCTION THIS MODEL OF
ROAD
LANE DETECTION IS IMPORTANT..
OBJECTIVE

To build a self driving car that knows how to drive in simulated environment.
The Car will be able to detect Lanes on the road with the help of lane lines.
To build an autonomous car which can control the steering angle.
TOOLS REQUIRED

Open-CV
Jupyter notebooks
Tensor flow object detection
Neural Networks
Python Libraries for machine learning
ASSUMPTIONS

CAR IS SELF DRIVEN


IT IS JUST A WORKING MODEL
IT WILL BEEP ONCE IT TOUCHES OR CROSSES THE LANE
IT MAY OR MAY NOT BR FOR REAL WORLD
Use Case Diagram-:
What we did in OpenCV

Reading the image


Displaying the image
Writing the image in directory
Reading the video
Changing the color space of image/video.
Gaussian blur on image.
Edge detection on image
Hough transform
OUR FIRST TASK THROUGH OPEN-
CV
Lane detection.
There are five steps to be performed to identify lanes
1. Changing to grayscale.
2. Applying Gaussian blur.
3. Applying Canny Edge Detection.
4. Apply masking to remove the non usable part of image.
5. Hough Line detection algorithm.
AGENTS USED
(Gray Scaling the image)
Need for converting the image to grayscale is to bring the image from three channels
to one channel.
Function used cv2.cvtColor(input_image,cv2.COLOR_BGR2GRAY)
Gaussian Blur
Need for applying the Gaussian blur on the image is to smoothen the image.
Function : cv2.GaussianBlur(img,(5,5),2)
Canny Edge Detection
This function is used to find all the edges in the image
Function used : edges = cv2.Canny(blur1,100,200)
Why used blur ?
If canny edge detection was applied on original grayscale image then we can get lots of
noises in the image.
So to remove the noises from the image we use the Gaussian Blur.
We have to remove the noise the clearly identify lane lines.
Masking
To correctly identify the lane lines we need to take that part of the image which
contain lane lines.
Remove the unwanted part by masking
Masking is done by bitwise and operator on two images.
How Masking is applied in previous case?
MY LEARNING

SOME OF THE THINGS WE LEARNT FROM THIS PROJECT ARE


1. IMAGE PROCESSING TECHNIQUES
2. MACHINE LEARNING MODELS
3. REAL TIME PROCESSING
4. INTEGRATION OF THE PARTS.

You might also like