• DocumentCode
    252310
  • Title

    Inverse Perspective Mapping based road curvature estimation

  • Author

    Dongwook Seo ; Kanghyun Jo

  • Author_Institution
    Dept. of Electr. Eng., Univ. of Ulsan, Ulsan, South Korea
  • fYear
    2014
  • fDate
    13-15 Dec. 2014
  • Firstpage
    480
  • Lastpage
    483
  • Abstract
    This paper proposes a solution for road curvature estimation. This information is relevant for the task of autonomous navigation. The proposed method is based on the Inverse Perspective Mapping (IPM) which gets rid of perspective distortion. From the IPM with multi scale lane markers appear as the parallel lines. Then, the lane markers are detected by using color segmentation, Hough transform and RANSAC. The detected lines are used for calculating the road curvature which is used to estimate a vehicle trajectory for autonomous vehicle or driver assistance system. In the current experiments, the correct detection rate reached 88.45%.
  • Keywords
    Hough transforms; driver information systems; image colour analysis; image segmentation; Hough transform; IPM; RANSAC; autonomous navigation; autonomous vehicle; color segmentation; driver assistance system; inverse perspective mapping; road curvature estimation; Estimation; Image color analysis; Image segmentation; Roads; Splines (mathematics); Transforms; Vehicles;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    System Integration (SII), 2014 IEEE/SICE International Symposium on
  • Conference_Location
    Tokyo
  • Print_ISBN
    978-1-4799-6942-5
  • Type

    conf

  • DOI
    10.1109/SII.2014.7028086
  • Filename
    7028086