| 欧氏聚类算法支持下的点云数据分割 |
| |
| 引用本文： | 陈向阳,杨洋,向云飞.欧氏聚类算法支持下的点云数据分割[J].测绘通报,2017,0(11):27. |
| |
| 作者姓名： | 陈向阳 杨洋 向云飞 |
| |
| 作者单位： | 1. 南通职业大学建筑工程学院, 江苏 南通 226007;2. 上海华测导航技术股份有限公司, 上海 201702;3. 河海大学, 江苏 南京 211100 |
| |
| 基金项目： | 国家自然科学基金（41174002） |
| |
| 摘 要： | 欧氏聚类算法是多元统计中的一种重要分类方法，可以将其应用于测绘领域中点云数据的分割。本文首先计算点云数据中两点之间的欧氏距离，将距离小于指定阈值作为分为一类的判定准则；然后迭代计算，直至所有的类间距大于指定阈值，完成欧氏聚类分割。具体步骤为：①利用Octree法建立点云数据拓扑组织结构；②对每个点进行k近邻搜索，计算该点与k个邻近点之间的欧氏距离，最小归为一类；③设置一定的阈值，对步骤②迭代计算，直至所有类与类之间的距离大于指定阈值。试验证明，欧氏聚类算法对不同测量技术手段获取的点云数据均具有适用性，可以成功对点云数据进行分割，分割效果良好。
|
| 关 键 词： | 欧氏聚类 点云数据 分割 算法 |
| 收稿时间： | 2017-07-26 |
Measurement of Point Cloud Data Segmentation Based on Euclidean Clustering Algorithm |
| |
| CHEN Xiangyang,YANG Yang,XIANG Yunfei.Measurement of Point Cloud Data Segmentation Based on Euclidean Clustering Algorithm[J].Bulletin of Surveying and Mapping,2017,0(11):27. |
| |
| Authors: | CHEN Xiangyang YANG Yang XIANG Yunfei |
| |
| Institution: | 1. Vocational College of Nantong, Nantong 226007, China;2. Shanghai Huace Navigation Technology Ltd., Shanghai 201702, China;3. Hohai University, Nanjing 211100, China |
| |
| Abstract: | Euclidean clustering algorithm is an important classification method of multivariate statistical analysis. It can be applied to the segmentation of point cloud in survey field. Euclidean clustering algorithm firstly calculates the Euclidean distance between two points. The points' distance which are less than a specified threshold are divided into a kind. Until all kinds of distance are greater than the specified threshold, it completes Euclidean clustering segmentation. Specific steps are as follows:① using Octree method set up topology structure of point cloud data; ② for each point conduct K-nearest neighbor search, calculating the Euclidean distance between point and K neighboring points, the minimum is regard as one kind; ③ set a certain threshold, the iterative calculation ② steps until all distances between the classes are greater than the specified threshold. Experiments show that Euclidean clustering algorithm is available to point cloud data obtained from different measuring means. It can successfuly segment point cloud data and the effect is good. |
| |
| Keywords: | Euclidean clustering point cloud data segmentation algorithm |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《测绘通报》浏览原始摘要信息 |
| 点击此处可从《测绘通报》下载免费的PDF全文 |
|
