高速公路车道线检测与跟踪算法研究

为提高高速公路上车道线识别的快速性和鲁棒性,提出了一种有效的车道线检测与跟踪方法。采用霍夫变换进行车道线检测,具有较强的抗干扰能力,能够准确地识别车道线。车道线跟踪利用Kalman预测参数建立感兴趣区域,然后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。由于搜索限制在预测范围内,提高了搜索精度,减少了搜索范围,保证了实时性能,且对虚线车道线识别特别有效。仿真实验结果表明,对于不同的天气状况和车道线种类,该算法均有较好的识别效果。

Research ON Lane Detection and Tracking Algorithm of Highway

An efficient method of lane detection and tracking is proposed to improve robustness and efficiency of lane recognition of highway.Lane detection based on Hough Transform(HT) has good robustness and can identify lane accurately.In lane tracking,the Region of Interest(ROI) is established using parameters predicted through the Kalman predictor.Then lane boundary points are searched by scan line method in ROI.The lane boundary in discontinuous area is located using Kalman predictors.The algorithm can improve searching precision,reduce the search area,and ensure real-time performance due to restricted searching area,especially for dashed lanes.Simulated experiment results show that the algorithm has good performance under different weather conditions and lane models.