主办:陕西省汽车工程学会
ISSN 1671-7988  CN 61-1394/TH
创刊:1976年

汽车实用技术 ›› 2021, Vol. 46 ›› Issue (15): 51-52.DOI: 10.16638/j.cnki.1671-7988.2021.015.014

• 智能网联汽车 • 上一篇    下一篇

一种基于视觉的车道线检测及追踪方法

高越, 张勇, 杜学峰, 杨伟   

  1. 长安大学 汽车学院,陕西 西安 710064
  • 出版日期:2021-08-15 发布日期:2021-08-20
  • 作者简介:高越,就读于长安大学汽车学院,研究方向:智能网联车及无人驾驶。

A Method of Lane Line Detection and Tracking Based on Vision

GAO Yue, ZHANG Yong, DU Xuefeng, YANG Wei   

  1. Chang'an University, College of Automobile, Shaanxi Xi'an 710064
  • Online:2021-08-15 Published:2021-08-20

摘要: 文章以OpenCV软件为主要平台,基于机器视觉建立一种车道线的检测及跟踪方法。首先检测出图片中车道线区域,对每一帧图片做透视转换到鸟瞰图视角。然后对鸟瞰图二值化,进一步区分左右两条车道线。用滑动窗口的方法检测第一帧的车道线像素点,从第一帧的曲线周围寻找后面的车道线像素点,拟合车道线曲线。经实验证明该方法能够准确地检测出车道线。

关键词: 机器视觉, 车道线, 滑动窗口, 检测, 追踪

Abstract: This paper uses OpenCV software as the main platform, and establishes a method of lane line detection and tracking based on machine vision. First, detect the lane line area in the picture, and convert each frame of the picture to the bird's-eye view perspective. Then binarize the bird's-eye view to further distinguish the left and right lanes. The sliding window method is used to detect the lane line pixels of the first frame, and find the subsequent lane line pixels around the curve of the first frame to fit the lane line curve. Experiments show that this method can accurately detect lane lines.

Key words: Machine vision, Lane line, Sliding window, Detection, Tracking

中图分类号: