基于形态学结构元素建模的车道线检测算法

针对智能车辆辅助导航系统中车道线检测算法复杂、鲁棒性差的问题，提出了一种基于形态学结构元素建模的车道线检测算法。算法首先通过交变序列滤波器对原始图像进行滤波，然后利用具有车道模型特征的结构元素对预处理后的图像进行特定目标提取，再对提取出的车道左、右标识线进行外线判断运算以去除双边缘，将所得结果进行整合得到车道左右内侧边缘线，最后进行Hough变换标记车道线。实验表明，该算法简单、鲁棒性强，能准确地检测到车道标识线。

Lane detection algorithm based on morphological structure-elements model

This lane detection algorithm based on structure-elements model of morphological simplified the complex processes of lane detection and improved robustness in intelligent vehicle vision assisted. In the algorithm, original image was filtered by alter sequence filter and special objects were extracted by the structure elements, which had similar characteristics to that of lane model. After that, the outside line of left and right lane was removed by a given means of estimation outside line. Finally, the resulted image was amalgamated and the lane mask was detected by means of Hough transformation. The experiments show that this algorithm is simple, of better robustness, and at the same time, it can efficiently detect the lane mask accurately.