摘要
针对存在移动车辆遮挡等强干扰的情况,提出了一种基于轮廓筛选的车道线检测方法。首先基于轮廓角度和像素数目滤除非车道轮廓进而实现特征点粗提取,然后提取像素点的位置信息,通过设计一种基于相对距离与梯度的自适应双阈值筛选法进行车道线特征点的精提取,进而拟合车道线。实验结果表明,该算法具有较强的抗遮挡能力,且满足实时性要求。
A lane line detection algorithm based on contour filtering was proposed for the problem of strong interference such as mobile vehicle occlusion.Firstly,the non-lane-contour was filtered based on the contour angle and the number of pixels.Then,an adaptive double threshold filtering method based on relative distance and gradient was designed to extract the feature points of lane line.Finally,the lane line was fitted.Experimental results show that the proposed algorithm has strong anti-occlusion ability and meets the real-time requirements.
作者
赵岩
赵建国
ZHAO Yan;ZHAO Jian-guo(School of Mechanical and Power Engineering,Zhengzhou University,Zhengzhou 450052,China)
出处
《科学技术与工程》
北大核心
2021年第7期2782-2787,共6页
Science Technology and Engineering
关键词
车道线检测
图像预处理
轮廓筛选
相对距离与梯度
lane detection
image preprocessing
contour filtering
relative distance and gradient
作者简介
第一作者:赵岩(1996-),男,汉族,河南新乡人,硕士研究生。研究方向:数字图像处理与模式识别。E-mail:1417312855@qq.com;通信作者:赵建国(1964-),男,汉族,河南周口人,硕士,副教授。研究方向:产品数字化设计、装配理论及应用、图形图像处理技术与图学理论及应用。E-mail:zzz8497@zzu.edu.cn。
