Skip to content

Real-Time Stereo Vision-Based Lane Detection System

Research output: Contribution to journalArticle

Original languageEnglish
JournalMeasurement Science and Technology
Issue number7
Early online date24 May 2018
DateAccepted/In press - 11 Apr 2018
DateE-pub ahead of print - 24 May 2018
DatePublished (current) - Jul 2018


The detection of multiple curved lane markings on a non-flat road surface is still a challenging task for vehicular systems. To make an improvement, depth information can be used to enhance the robustness of the lane detection systems. In this paper, a proposed lane detection system is developed from our previous work where the estimation of the dense vanishing point is further improved using the disparity information. However, the outliers in the least squares fitting severely affect the accuracy when estimating the vanishing point. Therefore, in this paper we use random sample consensus to update the parameters of the road model iteratively until the percentage of the inliers exceeds our pre-set threshold. This significantly helps the system to overcome some suddenly changing conditions. Furthermore, we propose a novel lane position validation approach which computes the energy of each possible solution and selects all satisfying lane positions for visualisation. The proposed system is implemented on a heterogeneous system which consists of an Intel Core i7-4720HQ CPU and an NVIDIA GTX 970M GPU. A processing speed of 143 fps has been achieved, which is over 38 times faster than our previous work. Moreover, in order to evaluate the detection precision, we tested 2495 frames including 5361 lanes. It is shown that the overall successful detection rate is increased from 98.7% to 99.5%.

Download statistics

No data available



  • Full-text PDF (accepted author manuscript)

    Rights statement: This is the author accepted manuscript (AAM). The final published version (version of record) is available online via IOP at Please refer to any applicable terms of use of the publisher.

    Accepted author manuscript, 23 MB, PDF document


View research connections

Related faculties, schools or groups