CASIA OpenIR  > 模式识别国家重点实验室  > 多媒体计算与图形学
局部形状特征概率混合的半自动三维点云分类
李红军; 刘欣莹; 张晓鹏; 严冬明
Source Publication浙 江 大 学 学 报(理学版)
2017-01
Volume44Issue:1Pages:1-9
Other Abstract
Pointcloudscapturedbythreedimensionalscannerhavebeenusedinmanyfields,includingmodelingof
digitalcities,acquisitionofthreedimensionalshapes,sceneanalysisandobjectmeasuring.However,duetothe
limitationofthesamplingprocessandthecomplexityofscannedscenes,mosttraditionalmethodsofsurfacemodelG
ingandthreedimensionalspaceanalysiscannotworkeffectivelywhendealingwiththepointclouddata.ClassificaG
tionisthereforeanimportantwayforpointcloudpreprocess.Fourfeatures,namelythevolumeofatetrahedron
constructedby4neighboringpoints,thedeviationofnormaldirectionsofneighboringpoints,thedeviationofprinG
cipaldirectionsofneighboringpoints,andthevaluesofprincipalcurvature,aremixedwithprobabilitiesforsemiG
automaticclassificationofthethreedimensionalpointclouddata.Withthenewmethod,apointcloudistobedividG
edintothreeclasses:planepoints,cylinderpointsandotherpoints.TheinitialclassificationresultislabeledaccordG
ingtoitssingleshapefeaturevalue.Theprobabilitymixtureiscompletedbyestimatingtheprobabilityofinferring
ashapefromalocalpointsetwithrespecttoeachfeature,generatingamixturewithweightedsum,andmaximizing
themixtureprobabilityfunction,whiletheprobabilityisestimatedwiththeaveragedistancebetweenapointandits
neighborpointstogetherwiththeconsistencyratioofinitiallabelsofthepointtoitsneighbors.Userinteractionsare
invokedtomakethechoiceofclassificationthresholdsandthesettingofweights,whichishelpfulwhendealingwith
pointcloudwithdifferentspacescaleandscanningpointresolution.Experimentsshowthattheproposedmethod
workswellforvariouskindsofpointclouddatasets,includingpointcloudsgeneratedbysimulation,andthosecorG
respondingtoasinglepinetree,astreetscene,acountryscene,andanairbornebigscene.
Keyword分类
Document Type期刊论文
Identifierhttp://ir.ia.ac.cn/handle/173211/14032
Collection模式识别国家重点实验室_多媒体计算与图形学
Recommended Citation
GB/T 7714
李红军,刘欣莹,张晓鹏,等. 局部形状特征概率混合的半自动三维点云分类[J]. 浙 江 大 学 学 报(理学版),2017,44(1):1-9.
APA 李红军,刘欣莹,张晓鹏,&严冬明.(2017).局部形状特征概率混合的半自动三维点云分类.浙 江 大 学 学 报(理学版),44(1),1-9.
MLA 李红军,et al."局部形状特征概率混合的半自动三维点云分类".浙 江 大 学 学 报(理学版) 44.1(2017):1-9.
Files in This Item: Download All
File Name/Size DocType Version Access License
局部形状特征概率混合的半自动三维点云分类(7217KB)期刊论文作者接受稿开放获取CC BY-NC-SAView Download
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[李红军]'s Articles
[刘欣莹]'s Articles
[张晓鹏]'s Articles
Baidu academic
Similar articles in Baidu academic
[李红军]'s Articles
[刘欣莹]'s Articles
[张晓鹏]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[李红军]'s Articles
[刘欣莹]'s Articles
[张晓鹏]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: 局部形状特征概率混合的半自动三维点云分类.pdf
Format: Adobe PDF
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.