基于车载激光点云数据的城市道路面提取方法研究

Research on Urban Road Surface Extraction Method Based on Vehicle-borne Laser Point Cloud Data

随着智慧城市建设的不断发展以及城市交通管理难度的日益增加,准确获取道路信息的研究已成为热点。本文针对城市结构化道路的空间特征,提出了一种基于车载激光扫描点云数据的道路面提取方法。该方法实现道路面提取的有效步骤为:首先对原始车载点云数据进行抽稀、降噪等预处理,使用一种布料模拟滤波(Cloth Simulation Filter, CSF)算法对预处理后点云数据进行滤波,去除非地面点对后续算法的影响;其次,使用主成分分析法估算得到各地面点局部法向量与曲率值;最后,使用改进的区域生长算法,结合各参数设置实现道路面点云的精确提取。通过两段实测数据对本文提出方法进行检验,实验结果的完整性r、准确性p以及检测质量q均在94%以上,表明在不同点云场景下,使用本文方法均能得到良好的提取结果,具有较高的适应性。

With the continuous development of smart city construction and the increasing difficulty of urban traffic management,the re-search on accurate acquisition of road information has become a hot spot.According to the spatial characteristics of urban structured roads,a road surface extraction method based on vehicle-borne laser scanning point cloud data is proposed in this paper.The effective steps of road surface extraction by this method are as follows:Firstly,the original vehicle-borne point cloud data is pre-processed such as thinning and noise reduction,and a cloth simulation filter(CSF)algorithm is used to filter the pre-processed point cloud data to remove the influence of non ground points on the subsequent algorithms;Secondly,the local normal vector and curvature of each surface point are estimated by principal component analysis;Finally,the improved region growth algorithm is used to extract the road point cloud accurately.The proposed method is tested by two sections of measured data.The integrity r,accuracy p and detection quality q of the experimental results are more than 94%,which shows that in different point cloud scenes,the proposed method can obtain good extraction results and has high adaptability.