车载LiDAR点云中的建筑物特征提取

Extracting Building’s Feature Based on Vehicle Borne LiDAR Point Cloud

针对车载LiDAR点云数据处理复杂、时间长的问题,本文以地物不同特征值作为建筑物自动提取算法的依据,通过点云数据预处理、聚类分析等一系列流程最终实现一般建筑物点云的自动提取。通过两个实验区点云数据的提取与相应的实际地物进行精度分析对比,结果表明本文算法对实例测区环境下的不同建筑物点云提取具有较好的有效性,满足数字城市三维建模的精度要求。

To solve the problem of complex and long time data processing in vehicle LiDAR point cloud,the features of ground objects are used as the basis of building automatic extraction algorithm in this paper.Through the pretreatment,clustering analysis and a series of processes of point cloud,it finally achieved the general building point cloud automatic extraction.It carried out experiments in two experimental area and compared the test accuracy with practical accuracy.The result shown that this algorithm had a good effect on the point cloud extraction of different buildings under the environment of the example survey area and the method meet the requirement of 3D modeling of digital city.