一种结合模糊C均值聚类的车道标识线识别方法  

A Fuzzy C-Means Clustering Based Lane Marking Identification Approach

作　　者：闫旭琴[1] 王知学[1] 李建新[1] 成巍[1] 

机构地区：[1]山东省科学院自动化研究所山东省汽车电子技术重点实验室,山东济南250014

出　　处：《山东科学》2010年第3期28-31,共4页Shandong Science

基　　金：济南市高新区科技攻关项目(济高管字[2008]232号)

摘　　要：在道路图像中,为了得到较理想的车道标识线的边缘,设计了一种车道标识线识别方法。利用图像中的边界信息、车道标识线边界的角度和灰度等特征,结合模糊C均值聚类提取车道标识线的边界。利用Hough变换、车道标识线边界的间断性实现车道标识线定位。实验结果表明该方法具有较好的准确性。This paper devises a lane marking identification approach in order to obtain the expected edge of a lane marking in a road image. The approach employs the edge information of a road image, angle and gray information of the edge of a lane marking and fuzzy c-means clustering to extract the edge of a lane marking. The approach then localizes a lane marking with Hough transform and the intermittent of the edge of a lane marking. Experimental result shows that the approach has better accuracy.

关 键 词：车道标识线识别 模糊C均值聚类 HOUGH变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]