Lane Recognition Technique Using Hough Transform

Abstract

In India, the rate of fatality due to accidents cases increases year by year. This problem becomes more serious and need all concerns to immediately take the following urgent actions in order to save precious human life. Therefore, this paper attempts to present a technique based on video and image processing for lane guidance warning system as an alternative way to reduce the accident rate due to lane departure. This system is capable to identify two lane markings from the video captured through a webcam. Warning system will automatically 'switch on' once lane departure occurs either to the right or to the left side of the road. The input video will first undergo threshold process and binary conversion. Then, lane marking detection is done using Hough transform, Hough line detection and local maxima finder. Finally, a decision-making algorithm is applied in the system to detect lane departure. It is found that, this system is able to identify straight lane markers, highlighting both markers with colour lines and able to give appropriate warning when lane crossing happen. For future enhancement, distance measuring function should be applied in the system so that the distance from one vehicle in front with another one at the back can be measured automatically.