车载激光扫描技术可以快速获取地物表面的高精度三维信息，作为一种新的数据获取手段，已逐渐应用于地理信息系统产业中。将激光扫描数据分类是对地物进行特征提取以及建模的前提与关键。现今，针对车载激光扫描数据的分类方法还不成熟。根据城市各典型地物空间特征(激光点云在三维空间的高程、与邻近点的斜率，以及其在二维投影平面上的分布、密集程度等)，本文提出一种主要适用于城市激光扫描数据的地物分类方法。首先，综合考虑车载激光数据的采集特点、车行GPS轨迹以及扫描数据中同一扫描线上相邻激光点之间的斜率关系，提取出路面。其次，对于非路面的激光点云数据，先使用基于格网化与区域分割相结合的方法进行实体划分，再通过计算地物空间形状特征的几项统计指标(外包围盒、实体高度等)，对实体进行分类。最后，以海南三亚市某街道为研究区验证该方法的有效性。实验结果表明，使用该方法能成功地分出研究区的路面、建筑物、树木和路灯四类地物，并进一步在同类地物间分出不同实体。

The vehicle-borne laser scanning (VLS) technology can quickly acquire three-dimensional information of earth surface with high precision. As a new technique of data collection acquisition, it has been gradually applied to Geographic Information System (GIS) industry. The classification on laser scanning points (LSP) is the premise and key step to the feature extraction and the model building of urban objects. This paper presents a classification method mainly suitable for the urban’s typical objects. Firstly, it extracts the LSP of the ground surface through taking three items into account, including characteristics of data collection, the vehicle's GPS trajectory, and the slope relationship between the adjacent LSP in the same scanning line. Secondly, as for the remaining LSP, the object division should be completed first through the methods of Grid and Segmentation, and then performs the objects classification by computing several statistical indicators of the spatial characteristics. Finally, this paper takes the Sanya city as the study area, to verify the effectiveness of this method. The result shows that, this classification method successfully classifies the categories of the ground, buildings, trees and street lamps from the LSP.