摘要
车道线识别是智能汽车视觉导航系统必须实现的基本功能。本文提出了一种基于扫描线的车道线识别算法。算法中使用定向边缘抽取算子和包含自适应阈值的双阈值二值化方法对道路图像进行预处理,使用扫描和区域划分提取车道线像素,并使用分段直线段道路模型建立车道线方程。本算法在路面情况较好的结构化道路上的平均识别率为95%。
Lane recognition is the basic function that a visual intelligent vehicle guidance system must implement. This paper describes a sweep-line based lane recognition algorithm. In this algorithm, we used directional edge detection operators and bi-thresh-old binarization which include an adaptive threshold in its two thresholds to preprocess the road image, then used scanning and region division to extract the lane-pixels, and used the segmented linear model to establish the lane equations. When on a well-conditioned structured roads, the average recognition rate is 95%.
出处
《微计算机信息》
北大核心
2008年第16期244-246,共3页
Control & Automation
基金
国家自然科学基金(60504003)
关键词
车道线识别
双阈值二值化
自适应阈值
扫描
区域划分
Lane recognition
Bi-threshoid binarization
Adaptive threshold
Scanning
Region division
