1种基于平行直线对模型的车道检测方法

A Lane Detection Method Based on the Double Straight-line Model

针对智能车辆安全辅助驾驶系统中利用单目视觉进行车道识别的问题,提出了1种基于平行直线对模型的车道检测方法。该方法根据高速公路图像特征构建平行直线对模型,在此基础上先利用Hough变换提取直线,再由改进的级联Hough变换检测出平行直线对的消失点,最后通过消失点和先验信息来提取当前车道线。使用Matlab对高速公路上不同路段、不同光照情况、不同车辆干扰下共150幅道路图像进行实验,检测精度达88.6%,平均检测时间为0.24s。实验结果表明,这一方法在高速公路行驶环境下能较准确地检测出当前车道线,具有很好的光照适应性、抗车辆干扰性和一定的实时性。

Aiming at improving the accuracy of traditional way of using monocular vision to identify the running lane ,a new method of lane detection based on the double straight line model is proposed .According to the characteristics of highways ,the method firstly builds a double straight line model ,and extracts straight lines by Hough Transformation , then detect the vanishing point of parallel lines by an improved Hough Transformation ,and finally extracts current traffic lanes through vanishing point and prior information from an image .An experiment is conducted based on 150 road images from different road sections ,with different illumination levels and different vehicle interferences .It is found out from the study results that the extraction accuracy is 88 .6% and the average detection time is 0 .24 seconds .Experimental results show that the proposed method can detect highway lanes with a good accuracy and it is also good at adapting to different illumination levels ,reducing the impact of vehicle disturbance and processing data in real-time .