摘要
点云数据被广泛用于多种三维场景,深度学习凭借提取特征自动化、泛化能力强等优势在三维点云的应用领域快速发展,逐渐成为点云分类的主流研究方法。根据提取方式的不同,将现有算法归纳为传统方法以及深度学习算法。着重介绍基于深度学习的代表性方法和最新研究,总结其基本思想以及优缺点,对比分析主要方法的实验结果;展望深度学习在点云分类领域的未来工作以及研究发展方向。
Deep learning has developed rapidly in the application field of 3D point clouds with the advantages of automatic feature extraction and strong generalization ability.It has gradually become the mainstream research method of point cloud classification.According to the different extraction methods,the existing algorithms are summarized into traditional methods and deep learning algorithms.This paper emphasizes the introduction of representative methods and the latest research based on deep learning,summarizes its basic ideas,advantages and disadvantages.It compares and analyzes the experimental results of the main methods.Finally,it looks forward to the future work and research development of deep learning in the field of point cloud classification.
作者
王文曦
李乐林
WANG Wenxi;LI Lelin(School of Resource Environment and Safety Engineering,Hunan University of Science and Technology,Xiangtan,Hunan 411201,China;Hunan Provincial Key Laboratory of Geo-Information Engineering in Surveying,Mapping and Remote Sensing,Hunan University of Science and Technology,Xiangtan,Hunan 411201,China)
出处
《计算机工程与应用》
CSCD
北大核心
2022年第1期26-40,共15页
Computer Engineering and Applications
基金
国家自然科学基金(41401497)
湖南省科技计划重点研发项目(2015GK3027)
湖南省自然科学基金(2018JJ3158)。
关键词
图像处理
点云分类
深度学习
特征提取
卷积神经网络
image processing
point cloud classification
deep learning
feature extraction
convolutional neural network
