摘要
对车道偏离预警线图像的边缘优化检测,能够有效降低道路交通事故的概率。对车道偏离预警线的边缘检测,需要提取车道线图像边缘特征点,将离散的特征点归类为不同的直线段,完成对车道偏离预警线图像边缘的优化检测。传统方法先提取车道线图像颜色特征点,再给出车道线感兴趣区域,但忽略了将离散特征点进行归类,导致检测精度偏低。提出基于模糊逻辑的车道偏离预警中车道线图像边缘检测方法。融合于最大信息熵原则得到车道线和背景的信息熵,确定各个隶属度函数的参数值,计算出车道线灰度阈值,给出摄像头与地面坐标之间的转换关系,得到标记线的边缘分布特性,提取车道线图像边缘特征点,将离散的特征点归类为不同的直线段,在此基础上完成对车道偏离预警中车道线图像边缘检测。仿真证明,所提方法检测精度高,为提升车道偏离预警系统的运行质量奠定了基础。
A detection method for image edge of lane departure from warning lane line is proposed based on fuzzy logic. Firstly ,the information entropy of lane line and background are obtained integrated with principle of maximum information entropy, and the parameter value of each membership function is confirmed. Then, the gray threshold of lane line is worked out, and transformational relation between camera and geographical coordinates is provided. Moreover, distribution character of edge of scratch line is obtained, and feature point of the image edge is extracted. The discrete feature point is classified into different line segments. Finally, the detection of image edge is completed based on that. Simulation proves that the method has high detection precision. It lays foundation of improving operation quality of warning system of lane departure.
作者
陈卫卫
王卫星
CHEN Wei - wei WANG Wei - xing(School of Information Engineering, Chang＇an University, Shanxi Xi＇an 710064, China)
出处
《计算机仿真》
北大核心
2017年第6期394-397,共4页
Computer Simulation
关键词
车道偏离
预警线
边缘检测
Lane departure
Warning line
Edge detection
