车载点云数据高速公路路面提取方法

Highway Road Surface Extraction Method Based on Mobile LiDAR Point Cloud Data

利用车载激光扫描数据提取高速公路的道路边界时,不能以路沿作为边界。针对此问题,提出了一种车载点云数据高速公路路面提取方法。根据车载激光扫描数据的特点,设计了逐步提取到精细优化的处理流程。首先,利用改进的RANSAC算法,在条件约束下进行分区自适应阈值粗提取;然后,通过结合RANSAC阈值的形态学滤波和基于点云密度和强度的滤波,进一步进行道路精细提取;最后,利用边缘检测算法对道路边界进行优化,从而提取完整道路点云。利用实际点云数据进行实验,结果表明,该方法提取的道路点云的准确度、完整度、提取质量均超过90%。

In highway mobile laser scanning point cloud data,it cannot take the road edge as the road boundary.Aiming at the problem,a highway filter extraction method is proposed.According to the characteristics of the mobile laser scanning data,the method designs a process which can gradually realize the fine highway road surface extraction.Firstly,the improved RANSAC algorithm is used for rough extraction of partitioned adaptive threshold under condition constraints.Then,the morphological filtering combined with RANSAC threshold and the filtering based on the density and intensity of point cloud is used to further fine extract the road.Finally,the edge detection algorithm is used to optimize the road boundary so as to extract the complete road point cloud.Experiments with actual point cloud data show that the accuracy,completeness and extraction quality of road point clouds extracted by this method are all over 90%.