基于改进区域生长法和霍夫变换的车道分割法

LANE SEGMENTATION METHOD BASED ON IMPROVED REGION GROWING AND HOUGH TRANSFORM

车道分割在智能交通监控系统中起着基础而重要的作用。精确地进行车道分割是车流量统计、车速测量等诸多智能交通应用的前提。以往基于种子区域生长法的车道分割方法对于较新且干燥的道路有较好的分割效果,但受地面上水渍、污渍、斑驳的影响较大,鲁棒性不强。为了解决这个问题,提出一种基于改进的种子区域生长法和霍夫变换的车道分割方法。通过改进的区域生长法,可以有效降低水渍、污渍、斑驳的干扰,提高系统的鲁棒性。

Lane segmentation plays a fundamental and crucial role in intelligent transportation monitoring systems.It is prerequisite for various intelligent transportation applications such as traffic volume collection and vehicle speed measurement to precisely segment lanes.Previous lane segmentation method based on seed region growing method works well on relatively new and dry roads,but is not robust enough to counteract the influence of water stains,dirt splotches or mottles.on the road surface.To tackle the problem,the paper proposes a lane segmentation method based on improved seed region growing method and Hough transform.The improved region growing method can effectively suppress the impact from water stains,dirt splotches and mottles;therefore the robustness of the system is improved.