Review: Lane Detection for Autonomous Vehicles Using Image Processing Techniques

Abstract

Lane line detection is given great attention in the study on autonomous driving and driver support systems, which, in the last few years, has become a crucial area in the field of intelligent transportation. Without any human involvement, autonomous vehicles can genuinely assess their surroundings, navigate, and provide transportation for people. The transportation system of the entire world will soon be replaced by autonomous vehicles. In order to accomplish this goal, the automobile industries are now researching in this area to maximize the benefits and address the issues [1]. Image processing is a major aspect of the electronic industry’s automation, protection, and surveillance-related applications. This study examined a detailed literature on autonomous vehicle systems, including various types of image preprocessing techniques, lane detection and tracking techniques, and these techniques are utilized to improve lane edge detection. Along with the Canny edge detection methodology, a cutting-edge method known as the Spiking Neural Network for lane boundary detection is also described in this paper. In order to speed up processing, an on-board camera of a vehicle initially sets the region of interest (ROI) on the original picture. The ROI is next subjected to picture preprocessing, which includes converting RGB to grayscale, stretching the grayscale, and using a median filter to remove noise. Hough transform is employed to identify the lane in this study work. Research findings demonstrate that such a strategy is much more reliable and accurate than alternative approaches.