基于单目视觉的车辆碰撞预警系统  

Monocular vision-based vehicle collision warning system

作　　者：刘志强[1] 温华[1] 

机构地区：[1]江苏大学汽车与交通工程学院,江苏镇江212013

出　　处：《计算机应用》2007年第8期2056-2058,共3页journal of Computer Applications

基　　金：江苏省科技厅2005年高新技术基金资助项目(BG2005028)

摘　　要：基于单目视觉的车辆碰撞预警系统能够发现道路前方的车辆并估算出与前方车辆之间的距离,利用预警机制及时提醒驾驶员危险状况。车道检测和车辆识别是该系统需要解决的两个主要难题,提出了利用边缘分布函数EDF检测车道标线,利用车辆底部纹理和对称性特征识别车辆,并根据图像坐标系和世界坐标系之间的几何映射关系测距。实验结果表明,提出的方法能够有效检测出车道标线,并能很好地测定与前车的距离。Monocular vision-based vehicle collision warning system can detect the front vehicle and estimate the distance between the host vehicle and the front vehicle, by providing warning mechanism it can protect drivers from danger in time. Lane detection and vehicle recognition are the major two problems in need to be solved in this research. Lane markings were detected using edge distribution function （EDF） and vehicle detection was achieved by using features, such as texture of underneath and symmetry property. The distance to front vehicle was then estimated from relation between image coordinate system and world coordinate system. The experimental results show that the proposed method can detect lane markings efficiently and the estimated distance is well provided to drivers.

关 键 词：碰撞预警 单目视觉 车道检测 车辆识别 

分 类 号：U467.14[机械工程—车辆工程]