It is a presentation for initial review of the project "Lane Detection". This project is useful for advanced driver assistance systems. We are developing this project by using computer vision. It includes gray scale conversion, noise reduction, canny edge detection, hough lane transform and some other user defined functions. The language we are using is python. Gray scale conversion converts the image from RGB format to gray. Since working with single colored channel image is much easier than working with three colored channel image. By using gaussian filter, noise reduction is performed. All the unwanted data, outliers, noisy data are removed. Simply the image is blurred. Next is canny edge detection, in this method edges present in the image are detected. And next region of interest is considered and hough lane transform is performed to get lanes on the road image.
1. Mini Project
Under the guidance of:
Mr. T. Praveen
BATCH NO : 19PB08
Class - CSE-B
Year - 4th
G. Keerthana - 19P61A0576
G. Aishwarya - 19P61A0571
2. Problem Statement
Objective
Existing System
Proposed System
Methodology
Architecture Diagram
Coding screenshots
3. The annual increase in car ownerships has caused traffic safety to become an important
factor affecting the development of a city.
The frequent occurrence of traffic accidents is caused by subjective reasons related to the
driver such as drunk, fatigue and incorrect driving operations.
4. Our objective is:
To develop a Lane detection system based on computer vision, providing the accurate
position of the vehicle in the lane.
To improve the driving safety of autonomous driving.
To reduce the burden of handling vehicle for the driver.
5. Manual transmission is a type of transmission in which the gears are changed by a lever
operated by the driver of a vehicle.
Drivers have to use both feet and have to drive one handed which can be considered as a
difficult task.
So automatic transmission has arrived and Road Lane Detection System is one of the crucial
function in it.
6. The proposed system decreases the frequent occurrence of traffic accidents caused by
subjective reasons related to the driver, such as drunk, fatigue and incorrect driving
operations.
This system describes the path for self-driving cars by detecting the lanes on the road and
thus avoiding the risk of getting in another lane.
7. Following are the steps to be followed:
Capturing and decoding video file: We will capture the video using Video Capture object
and after the capturing has been initialized, every video frame is decoded (i.e. converting
into a sequence of frames).
Grayscale conversion of image: The frames in the video are in RGB format, RGB is
converted to grayscale because processing a single channel image is faster than processing a
three-channel colored image.
8. Reduce noise: Noise might cause misleading edges, picture smoothening is required before
proceeding. This procedure employs a Gaussian filter.
Canny Edge Detector: It is an edge detection operator which detects the wide range of
edges in images.
9. Region of interest: This stage involves merely considering the area covered by the road
lane. Here, a mask of the same size as our road picture is constructed. Furthermore,
between each pixel of our canny picture and this mask, a bitwise AND operation is done. It
eventually masks the canny picture and displays the region of interest outlined by the
mask's polygonal shape.
10. Hough Line Transform: The Hough Line Transform is a line detection transform. For this
we use Hough Line Probabilistic Here, the transform is utilised, and the result is the
extremes of the identified lines. All of these procedures are critical in detecting lanes. The
main assumption is that the lanes are long and parallel, allowing the Hough transformation
method with edge selection to find them.