在分析传统Hough变换优缺点后,提出一种基于CCD参数智能调节的车道线检测方法。通过建立图像特征区域,调节CCD的亮度、增益和曝光时间等参数,增强车道线与周围道路的对比度,减少路面噪声;在此基础上提出改进的Hough变换算法,对整幅图像进行种子点的选取与归类,再对每组种子点进行Hough变换,最后通过一定的角度约束提取车辆当前运行车道线。道路试验验证了该方法的有效性、实时性和准确性。
The strong and weak points of traditional Hough transform are analyzed and a novel lane mark detection scheme based on intelligent adjustment of CCD parameters is proposed. By establishing image feature regions and adjusting the parameters of brightness, gain and exposure time of CCD, the contrast between lane marks and road surface is enhanced and road image noise is reduced. On this basis, an improved Hough transform algorithm is put forward, and the selection and classification of seed points are conducted on the whole image, and the Hough transform is performed on each seed group. Finally the current lane marks are extracted through certain angle constraints. The results of road test verify the effectiveness, accurateness and real time performance of the proposed scheme.