影像密集匹配点云的单体化提取

发布时间：2018-03-23 02:09

  本文选题：倾斜影像　切入点：密集匹配点云　出处：《测绘通报》2016年12期 　论文类型：期刊论文

【摘要】：以倾斜影像密集匹配点云数据为研究对象,提出了一种快速提取建筑物单体点云模型的方法。首先通过过滤地面点云数据以实现地物分离,采用高程频率直方图确定植被点云的高程区间并予以剔除。然后采用基于密度的聚类分析算法对剩余建筑物顶部点云进行了聚类。最后根据每栋独立建筑物屋顶点云的坐标范围,从原始点云数据中提取了每栋独立建筑物的完整点云,实现了对于倾斜影像密集匹配点云的单体化提取。
[Abstract]:Based on the dense matching point cloud data of inclined image, a fast method for extracting the single point cloud model of buildings is proposed. Firstly, the ground point cloud data are filtered to realize the separation of ground objects. The elevation frequency histogram is used to determine the height interval of the point cloud of vegetation and to eliminate it. Then the cluster analysis algorithm based on density is used to cluster the point cloud at the top of the remaining building. Finally, according to the roof of each independent building, the point cloud is clustered. The coordinate range of the point cloud, The complete point cloud of each individual building was extracted from the original point cloud data, and the single point cloud was extracted from the dense matching point cloud of inclined image.
【作者单位】： 信息工程大学;
【分类号】：P23
，

本文编号：1651416