Lane marking recognition based on laser scanning

Towards safe autonomous vehicle navigation the problem of lane detection and classification is highly important in the development of advanced driver assistance system (ADAS). This paper proposes a new method to detect the road lane marking for safe autonomous navigation purpose. It focuses on unconventional methods of identifying lane markings on a road surface through Laser Measurement System (LMS). This method was executed in three steps. Firstly, to detect lane markings a Density-based spatial clustering of applications with noise (DBSCAN) method was implemented. Secondly, in order to determine the surface course a distance clustering analysis was proposed. Thirdly, the Random Sample Consensus (RANSAC) line fitting method was implemented for removing the noise points around the road lane area. Lastly, an automatic peak detection was implemented to perform lane marking detection on road surfaces. Preliminary results were performed and tested on a group of consecutive fames to prove its effectiveness.