A Vison Based Lane Detection Approach Using Vertical Lane Finder Method

This paper presents a novel approach for detection of lane using Kirsch operator and VLF method. The problem of detecting lane using vision sensors poses many problems such as atmosphere of the road, traffic intensity on the road, pollution, shadow of a tree, and other objects on the road. Normally, in any roads, the lanes are colored with yellow and white with different purpose. To highlight yellow and white lanes in a given RGB image, conversion from RGB to HSL and RGB to HSV is performed. The Gaussian filter is employed to smooth the image and to remove the noise, followed by edge detection using Kirsch operator. After edge detection, to retain only candidate lane lines in edge map, we compute the orientation of gradient and select the pixels which are representing candidate lanes. Finally, a VLF method is used to detect lanes in the image. The experiments are conducted using KITTI benchmark dataset and results are evaluated using popular metrics. Through experiments we demonstrate that detection of lane using proposed method yields promising accuracy.