基于机器视觉的快速车道线识别

为了克服已有车道线识别算法运算复杂、速度较慢以及鲁棒性欠缺等不足之处, 提出一种新的快速车道线识别算法, 首先通过对图像的灰度变化分析, 得出车道线轮廓像素, 然后运用B-Spline曲线拟合车道线轮廓, 得到最终的识别效果图。实验表明, 该算法在速度和识别率上都能取得优异的表现。在嵌入式平台上, 该算法取得了12 fps的速度, 符合智能驾驶的实际需求。

Computer vision based fast lane detection

In order to overcome shortcomings of previous lane detection algorithms in computational complexity, speed and robustness, this paper presented a new and fast lane detection algorithm. First, it detected lane edges by analyzing grey scale change of the images. Then it used B-Spline fitting to match the lane edge pixels to get final detection results. Experiments show that the proposed algorithm has better performance than previous works in both speed and effectiveness. On the embedded platform, the algorithm processes 12 fps. It can meet the practical needs of intelligent transportation.