基于Hough变换的车道线检测算法设计

Design of Lane Detection Algorithm Based on Hough Transform

结构化道路具有完整而明显的特征信息,改进了传统道路车道线检测算法,能够满足车道线检测的实时性和鲁棒性的要求。首先提取感兴趣区域,减少数据处理时间。采取自适应阈值的二值化处理,能够保留更多的车道线信息。根据先验知识,对于不符合实际几何特征的车道线进行剔除,得到最符合实际的车道线。通过对多种环境图像的实验,可以验证该算法能够满足大多数行车环境。

Structured roads have complete and obvious feature information.This paper improves the traditional lane detection algorithm,which can meet the real-time and robust requirements of lane detection.Firstly,the region of interest is extracted to reduce data processing time.The adaptive thresholding two-valued processing can retain more lane information.According to the prior knowledge,the lanes which do not conform to the actual geometric characteristics are removed,and the most realistic lanes are obtained.Through experiments on various environmental images,it can be verified that the algorithm can meet most of the driving environment.