基于双曲线模型的车道线检测算法

Algorithm for Lane Detection Based on Hyperbolic Model

针对车道线识别准确性、可靠性等问题,提出了一种基于双曲线模型的车道线检测算法。首先,对图像进行预处理,采用Sobel算子实现可靠的车道边缘检测;然后,运用Hough提取车道线边界信息,并运用区域生长法筛选道路边缘点;最后,结合车道线边界参数和双曲线模型参数,基于最小二乘法完成车道线重建。试验结果表明该算法可准确识别不同环境下的车道线。

Aiming at the accuracy and reliability of lane recognition,an algorithm for lane detection based on hyperbolic model was proposed.Firstly,the preprocessing of a road image was conducted with the method of Sobel operator.Then,the lane boundary information was extracted by the means of Hough,and the region growing method was applied to filter road edge points.Finally,the lane was re-constructed based on the least square method combining with the extracted parameters of lane boundary and hyperbolic model.The results show that the algorithm can identify lanes in different surroundings accurately.