基于分段切换模型的快速车道线检测

摘要

In order to make up for the shortcomings of the single model in lane line detection and improve the real-time and accuracy of the lane line recognition, a highway lane line detection algorithm combining Hough transform and quadratic curve model is proposed.First, the algorithm divides the region of interest.And on the basis of LOG operator for edge detection of the ROI, the Hough transform is performed to detect the straight line.Second, according to the detected line, it further dynamically divides ROI and straight line detection.Finally, the straight line and the quadratic curve are switched to determine the lane line by judging the direction change trend of the two straight segments.Through the experimental test of the actual lane of the freeway, the experimental results show that, the algorithm can accurately detect the lane line, the accuracy and practicality are higher than the contrast algorithm.%为了弥补单一模型在车道线检测中的缺点,提高车道线识别的实时性和精确性,提出一种由Hough变换与二次曲线模型相结合的高速公路车道线识别算法.该算法首先对图像作感兴趣区域ROI(Region of Interest)划分;再在LOG算子对ROI边缘检测的基础上进行Hough变换检测直线;然后根据已检测出的直线再进一步动态划分ROI和直线检测;最后通过判断两直线段的方向变化趋势进行直线和二次曲线切换拟合车道线.通过对高速公路实际车道线进行实验检测,实验结果表明,该算法能够准确地检测出车道线,准确率和实用性均高于对比算法.