Lane Detection Using Orthogonal Wavelet Transform and Polynomial Line Fitting

Lane detection can help to increase the efficiency of drivers. This paper presents the development of a road lane detection algorithm using image processing techniques. This process is developed based on static images. Some difficulties are faced during lane detection such as blurred image, excessive light condition. To detect the lane correctly avoiding these natural conditions is the main contribution of this paper. Lane detection using orthogonal wavelet transform and polynomial line fitting is implemented in this research. For preprocessing, extraction of the region of interest (ROI) is the main part and first step of this paper. Secondly, this paper proposes an adaptive threshold value to convert the region of interest image to binary after compressing the image. Thirdly, remove unwanted region based on angle. Then get lanes using a simple algorithm. Finally, to determine either left or right lane this paper proposes another simple algorithm. Experimental results indicate agreeable execution with a normal recognition rate of 95% under different light conditions. Additionally, recognition rate is increased compared with existing methods.