散乱点云数据的分割方法探究

Segmentation of Scattered Point Cloud Data

数据分割是点云数据处理流程中的一项关键技术,本文针对点云数据分割这一问题,分析了基于几何模型的分割法、欧几里德簇分割法和区域生长分割法三种点云数据分割方法的原理和算法流程,并利用实际工程数据检验了三种分割方法的效果。针对以上三种方法的优势及缺陷,本文实现了附条件的欧几里德簇分割法,该算法对欧几里德簇分割法和区域生长分割法进行融合改进,利用同一点云数据对该算法与上述算法进行比较分析,最后用一处复杂建筑物点云数据对该算法的分割效果进行再次检验,分割效果较为明显,再次证实了该算法具有较强的实用性。

Data segmentation is an important technology in the point cloud data processing. This paper focuses on the topic of point cloud data segmentation methods,analyses three segmentation methods based on the geometric model of the segmentation method,the Euclidean cluster segmentation method and region growing segmentation method,and then uses the actual engineering data to test the result of three segmentation methods. For the advantages and shortcomings of all above three methods,this paper improves integration of the Euclidean cluster segmentation method and region growing segmentation method,achieving the conditional Euclidean cluster segmentation method. Compared with other algorithms,the conditional Euclidean cluster segmentation method gets the more significant results.