基于动态区域规划的双模型车道线识别方法

Lane Detection System Based on Region Dynamically Planning and Dual Model Fitting

  • 摘要: 为提高车道偏离预警以及前方车辆识别系统性能,提出一种基于动态区域规划的双模型结构化道路车道线检测方法。通过基于模板的腐蚀、膨胀处理改善了图像预处理效果。根据“微元”理论进行图像动态细化分区,对各区域局部独立识别,整体综合分析,最终利用直线和B-样条曲线双模型完成车道线拟合。 实车试验证明,直线/B-样条曲线双模型拟合,提高了算法对不同车道线的适应性。

     

    Abstract: In order to improve the performance of highway lane departure warning system, a lane detection and tracking system was presented based on region dynamically planning and dual model fitting. The image pre-processing was featured by a method of template erosion and dilation to remove the interference information. According to the infinitesimal theory, the picture was divided into small pieces, which were extracted separately, and analyzed as a whole. Finally the straight line and B-spline models were applied to simulate the lane. Test results of this system show that the system and its arithmetic improve the adaptability to different lane markings.

     

/

返回文章
返回
Baidu
map