文献类型：期刊文献

中文题名：基于视频图像的车道线识别方法

英文题名：Approach for Lane Marking Recognition Based on Video Image

作者：张军平[1];殷勇辉[1]

机构：[1]华东理工大学机械与动力工程学院,上海200237

年份：2009

卷号：33

期号：S1

起止页码：144

中文期刊名：电视技术

外文期刊名：Video Engineering

收录：北大核心:【北大核心2008】；

语种：中文

中文关键词：车道线识别;形态学梯度;Hough变换

外文关键词：lane marking recognition;morphology gradient;Hough transform

摘要：提出了一种适用于结构化道路的车道线识别算法,通过使用形态学梯度算法来配合Hough变换,再根据道路图像特征对Hough变换进行条件约束。通过大量试验结果表明,利用本算法对连续的、断续的、磨损的车道线均能准确地识别出来,而且达到了15 f/s(帧/秒)的处理速度。
A way of lane marking recognition arithmetic being acclimated to structuring road is put forward.It makes use of morphology gradient to coordinate with Hough transform,and restrict Hough transform with road image feature.Through the results of many experiments,it can detect continuous,making and breaking,abrasion lane marking exactly at fifteen frames per second,so it is an effective arithmetic of lane marking recognition.