中文摘要：

通过介绍三维点云数据分割的基本原理和特征,以及经典的点云数据集和测试平台,总结、对比现阶段各类点云分割算法的基本原理、特点和适用场景,指出现阶段点云分割算法存在现有算法的自适应能力差、大部分分割算法对异常点和噪声敏感,并且算法的分割效率也有待提升等问题.未来的研究除需解决上述问题外,还应充分利用点云数据的语境信息,进一步结合深度学习理论,从而提升点云分割效果.

英文摘要：

The basic principles and characteristics of 3D point cloud data segmentation, classical point cloud data set and test platform were introduced. The existing problems were summarized by summarizing and contrasting the basic principles, characteristics and scenario： the adaptive ability of the existing algorithm is poor, most of the segmentation algorithm is sensitive to outliers and noise, and the segmentation efficiency of the algorithm should be improved. In addition to solving the above problems, the context information of the point cloud data should be fully exploited combined with further theory study to enhance the point cloud segmentation results.