基于图像的非接触式人体测量系统研究与实现

《基于图像的非接触式人体测量系统研究与实现》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、苏州大学硕士学位论文基于图像的非接触式人体测量系统研究与实现姓名：卢晨申请学位级别：硕士专业：计算机软件与理论指导教师：张广泉;王辉20090501AbstractNon—contactinghumanbodyauto—measurementtechnologyobtmnshumanbodysizequicklyandexactly,toestablishhumanbodyformdatabase．Itisappliedwidelynotonlyforthestudyofhumanbodymeasurement，thestandardofclothings

2、izeandcustom．made，humanbodymodelingforallkindsofperson，butalsoforprovidingtheresearchresourceaboutbodilyform仃endsofOurcountryanddatasustainingbywhichexpertsdesignmorecomfortable，morecustomizedlydailynecessities．Inourcountry，it’SverylateforthestudyofNon-contactinghumanbodyauto-meas

3、urement．WeintroduceintoadvancedequipmentsfromAmerican，Europe，andJapan．TheseequipmentsaretooexpensiveandnotfitforOurcountries’bodilyform．Meanwhile，thestudyisnotappliedforclothingmanufactureandstillforexperiments．Aimingatabovedisadvantages，wedevelopanon—contactinghumanboayauto。measu

4、rementsystembasedon2Dimages．Themainstudyinthispaperisthat：(1)Becauseofthedisadvantagethatwecan’tobtainthehumanbody’Scircumferencessizefrom2Dimages，wesortandanalysisforcircumferences，mappingtheexamples’circumferencecurvesapproximately．Thenwemakerelationshipusingfunctionalmodelbetwe

5、enthecircumferencesanditswidth，depth，humanstatureandavoirdupois．Thus，itiseasytoestimatethesizesbythemodels．Otherwise，wepresentaapproachaboutbasedexamplefunctionalmodelofhumantrunkforactualhumanbodymodeling．(2)Aimingatthecontourpointscontainlessinformationexcepttheircoordinateswhic

6、hislowlevelforimages．Soweneedrecongnisesizepointsbysearchingallcontourandit’Slowrecongnitionratio．Wepresentamethodtodividehumanbodycontourbasedorientation．Themethodaddsthegeometryrelationshipattributeandobjectrelationshipsttribrtetodividecontourpoints．Thencombiningtwodivisions，wec

7、anobtainthedivisionofwholebodycontourbytheknowledgeofsettheory．Sotheinformationcontainedinthecontourpointsisextendedfromdescriptionofhumancontour．Thethelow-levelcoordinatedescriptiontohigh．1evelmethodcallextractthespecificpartsoftheprofile，HTheStudyandImplementationofImage-basedNo

8、n—contactingAutomaticBodyMeasurem