一种用于高速公路的快速有效的车道线识别算法

提出了一种针对高速公路的车道线检测与跟踪方法;在图像预处理中采用基于采样的自适应阈值以满足不同光照条件下的使用要求,采用霍夫变换(HT,Hough Transform)进行车道线初始检测,车道线跟踪利用Kalman预测参数动态建立感兴趣区域(ROI,Region of Interest),用扫描线法搜索车道线边界点,在车道线间断区域利用Kalman预测器定位车道线边界;设计了一个失效判别模块,验证跟踪结果,当跟踪失败时,重新启动初始检测算法进行识别;实验结果表明,对于不同的车道线种类和在大部分车道线被前方车辆遮挡的条件下,该算法均具有较高的实时性和鲁棒性。

A Fast and Efficient Lane Recognition Algorithm for Highway

A method of lane detection and tracking for highway is proposed.The adaptive threshold based on sampling was used for different illumination demands during image pretreatment.Lane was initially detected by Hough Transform.In lane tracking,the region of interest(ROI) was established with parameters predicted through the Kalman predictor,the lane boundary points were searched by scanning line method in ROI,and the lane boundary in discontinuous area was located by Kalman predictor.A judgment module that was responsible for estimating tracking results was introduced.If tracking algorithm was disabled,initial detection algorithm was restarted.For various kinds of lanes and most lanes covered with vehicles ahead,experiment results indicate that the algorithm has good robustness and efficiency.