Abstract: In order to extract road elements from point cloud data and provide a basis for the digitization of road infrastructure and the production of high-precision maps, an automatic extraction process is developed in this paper by combining ground segmentation and pavement extraction. Firstly, the ground segmentation is carried out, and the algorithm is used to preprocess the point cloud data to realize the effective separation of the ground point cloud and the non-ground point cloud. Secondly, the seed points are selected, and the seed points representing the typical pavement characteristics are selected based on the intelligent algorithm of the geometric characteristics of the point cloud. Finally, the regional growth algorithm is used to automatically extract the pavement, which solves the problem of over-segmentation of the growth algorithm.

Key words: point cloud date; machine learning; typical pavement characteristics; automatic extraction

摘要： 为了从点云数据中提取道路要素，为道路基础设施的数字化和高精度地图的制作提供 基础，文章结合地面分割和路面提取开发了一种自动化提取过程。首先进行地面分割，采用 算法对点云数据进行预处理，实现地面点云与非地面点云的有效分离；其次选取种子点，基 于点云几何特性的智能算法，选定代表典型路面特征的种子点；最后使用区域生长算法对路 面进行自动化提取，解决了生长算法的过分割问题。

关键词: 点云数据；机器学习；典型路面特征；自动化提取