基于语义信息的机载LiDAR建筑物点云提取  

Building points extraction algorithm based on semantic information of airborne LiDAR data

作　　者：陈光洲 刘勇[1] 王亚芹[1] Chen Guangzhou;Liu Yong;Wang Yaqin(Zhejiang Tongji Vocational College of Science and Technology,Hangzhou 311231,China)

机构地区：[1]浙江同济科技职业学院,杭州311231

出　　处：《工程勘察》2018年第10期61-65,72,共6页Geotechnical Investigation & Surveying

摘　　要：本文研究了机载LiDAR数据中建筑物点云提取的算法。首先,分析典型地物点云的分布特征及规律,形成语义信息;利用TIN的数据组织形式,以三角形边长、面积、坡度、相邻三角形坡向等语义信息作为限制条件将植物点剔除;再利用高程分布直方图确定阈值,剔除地面点,进而得到建筑物点集。选取机载LiDAR数据进行实例验证,并将本文算法与其他两种算法进行了对比。结果表明,本文算法能较好地提取建筑物屋顶点集,精度高于内插类算法,提取效果优于未考虑语义信息的算法。In this paper, we study the algorithm of building points extraction using airborne LiDAR data. Firstly, the distribution characteristics of the typical feature points are analyzed and summarized, and the semantic information is formed. The semantic information, such as side length, area and slope of the triangle, aspect between adjacent triangles, are used as the limiting conditions to remove the points from the point of TIN. Then, the threshold is determined by the histogram of height distribution, and the ground points are removed, so as to get the building points. Airborne LiDAR data are selected for verification test. The algorithm is compared with the other two algorithms. Experiments show that this method can extract the roof points set of buildings and the accuracy is higher than the interpolation algorithm. The extracting result is better than the algorithm without considering semantic information.

关 键 词：语义信息 不规则三角网TIN 机载LIDAR 建筑物点云提取 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]