Lane-Finding Based on Structure Analysis of Lane and Computer Vision

Advanced driver-assisting systems (ADAS) performs a crucial role for the safety and ease of transportation. Finding of lane markings is an essential step of ADAS. For the protection of automated vehicles, the precise results of lane detection are compulsory. In this study, we present an innovative approach based on structural analysis of lane and Computer Vision Technique such as Hough Transform (HT) for the detection of lane markers. Firstly, a stage of preprocessing is applied to remove the pavement area that develops the lane markings background. Next, an edge detection technique is applied to generate the map of the edge and masking is performed to obtain edges in our region of interest (ROI). Then, using HT to find lanes, important lines are analyzed in the area of interest. In end, the concept of vanishing point used to predict lane has been explained. Experimental result states that the proposed system can find boundaries of the lane and exact lane departure alert in the existence of several image artefacts, such as light change, bad lane mark. The proposed technique displays more efficiency.