中文摘要：

车载激光扫描系统获取的结构化道路环境（城市道路/高速公路）点云数据量大、目标复杂,难以有效提取出道路的点云。本文通过分析扫描线上激光点云的空间分布和统计特征,提出一种适用于结构化道路环境的道路点云自动提取方法。在Optech公司提供的两份车载激光扫描点云数据中,道路提取结果的完整率、准确率、提取质量相应地超过94.92%、95.80%和91.13%。通过定量和定性的试验分析,该方法不仅适应于有固定道路宽度的规则道路提取,同样适用于无固定宽度的非规则道路提取。

英文摘要：

The mobile laser scanning acquires very dense point clouds along road corridors.Dense point clouds with high resolution and rich road-scene objects contribute the difficulties of road extraction.An automated method for extracting structural road from mobile laser scanning point clouds by the feature of scanning line with spatial distribution and statistics is proposed.Experiments were undertaken to evaluate the validities of the proposed method with two datasets acquired by Optech＇s Lynx Mobile Mapper System.The completeness of our road extraction results is over 94.92%,correctness is above 95.80%,and quality is above 91.13%,which proves a promising solution qualified for structural road with regular or irregular shape from mobile laser scanning point clouds.