中文摘要：

本文提出了一种建立动态搜索带的新的车道识别算法,使其能够更好的处理非理想路况的各种不确定因素,建立实时动态的搜索带,除了能更精确的追踪车道标志线之外,还能更好的应用于弯路路况。运用LMedSquare曲线拟合算法更好地提高了车道拟合过程的稳健性和精确性。由于搜索带的建立,整个识别过程的计算量大大降低,因此更适合实时要求。

英文摘要：

The paper brings forward a lane detection algorithm based on dynamic search zones,which deals with uncertainty on road better,and sets up real-time dynamic search zones.Besides more accuracy on lane-tracking,this algorithm also can be applied in a crooked road condition.The application of LMedSquare curve fitting improves stability and precision of the lane fitting process.Due to the search zone,computation of the search process can be reduced greatly,suitable for real-time applications.