Journal of Terahertz Science and Electronic Information Technology , Volume. 20, Issue 4, 372(2022)

Lane line tracking method combining frame difference method and window search

LI Chunming1、*, YANG Shan1, and YUAN Songling2
Author Affiliations
  • 1[in Chinese]
  • 2[in Chinese]
  • show less

    In the process of drawing a lane line map for a computer, it is necessary to accurately track the function requirements of the lane line in the moving background. A lane line tracking method is proposed by combining the frame difference method with the window search. Firstly, the image captured by the wide-angle lens is corrected by the checkerboard, and then the Region Of Interest(ROI) including the lane line is converted into a bird's-eye view by using the Inverse Perspective Mapping(IPM), and the white and yellow pixels are separately filtered by using Hue,Saturation,Value(HSV) and Red Green Blue(RGB). Secondly, the lane line is corrected with the angle between the lane line and the vertical line. According to the pixel density of the corrected picture, the starting point of the lane line is selected, and the entire lane line is extracted by the sliding window search method. Finally, the improved frame difference method is utilized to track the lane lines and standardize the lane line pixels according to the lane line standard. A large number of actual road driving tests show that the accuracy of this algorithm is 94.97%, which can accurately complete the tracking of lane lines.

    Tools

    Get Citation

    Copy Citation Text

    LI Chunming, YANG Shan, YUAN Songling. Lane line tracking method combining frame difference method and window search[J]. Journal of Terahertz Science and Electronic Information Technology , 2022, 20(4): 372

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category:

    Received: May. 25, 2020

    Accepted: --

    Published Online: Aug. 12, 2022

    The Author Email: Chunming LI (fugirlim@163.com)

    DOI:10.11805/tkyda2020231

    Topics