基于深度学习的大规模点云语义分割方法综述  

A survey of large-scale point cloud semantic segmentation based on deep learning

作　　者：双丰 黄兴文 李勇 刘修宁 覃建川 李旭 SHUANG Feng;HUANG Xingwen;LI Yong;LIU Xiuning;QIN Jianchuan;LI Xu(Guangxi Key Laboratory of Intelligent Control and Maintenance of Power Equipment,School of Electrical Engineering,Guangxi University,Nanning 530004,China;Artificial Intelligence Key Laboratory of Sichuan Province,Yibin,Sichuan 644000,China)

机构地区：[1]广西大学,电气工程学院广西电力装备智能控制与运维重点实验室,南宁530004 [2]人工智能四川省重点实验室,四川宜宾644000

出　　处：《测绘科学》2023年第2期195-209,共15页Science of Surveying and Mapping

基　　金：广西自然科学基金项目(2022GXNSFBA035661);人工智能四川省重点实验室开放基金项目(2021RYJ06);广西制造系统与先进制造技术重点实验室开放基金项目(20-065-40S005)。

摘　　要：为了展示深度学习在点云处理上最新进展,同时促进对点云语义分割方法的研究,该文对基于深度学习的大规模点云语义分割方法进行了综述。在介绍8个室内和室外语义分割数据集的基础上,重点对近几年的深度学习点云语义分割方法进行了归纳和分析,并在S3DIS、Semantic3D、Toronto3D、ISPRS Vaihingen 3D和SemanticKITTI数据集上对不同方法进行了比较,并构建了相应的基准。最后对目前点云语义分割算法存在的问题和未来趋势进行了分析。In order to provide the latest progress of deep learning in point cloud processing,and promote the research of point cloud semantic segmentation methods,a comprehensive review of semantic segmentation methods based on deep learning was sorted out in this paper.First,eight indoor and outdoor semantic segmentation datasets were introduced.Next,semantic segmentation methods of point cloud in recent years by deep learnning were summarized and analyzed.Then,different methods were compared on S3DIS,Semantic3D,Toronto3D,ISPRS Vaihingen 3Dand SemanticKITTI datasets,and corresponding benchmarks were constructed.Finally,the existing problems and future trends of point cloud semantic segmentation algorithms were analyzed.

关 键 词：深度学习 激光点云 语义分割 综述 

分 类 号：P237[天文地球—摄影测量与遥感]