欢迎来到天天文库
浏览记录
ID:36566739
大小:3.66 MB
页数:72页
时间:2019-05-12
《基于图像颜色的石材分类算法及测试平台研究与实现》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、武汉理工大学硕士学位论文基于图像颜色的石材分类算法及测试平台研究与实现姓名:康利娟申请学位级别:硕士专业:计算机应用技术指导教师:陈先桥20090601AbstractWiththerapiddevelopmentofinformationtechnologyandnetworktechnology,thepictureinformationdatabaseofhuge,differentcontentsisappearingconstantly.AndthepictureresourcesespeciallyonInternet,aredoublingandredou
2、blingconstantly.Asthesametime,thedevelopmentofstoneindustryleadstothegrowthofthestone.imageinformationdatabase.Inordertoretrievetheneededimagerapidlyande伍ciently,theContent-basedImageRetrieval(CBm)hasemergedin20mcentury90’S.CBIRanalysestheimageonthebasisofthecoloLtextureandshapeembeddedi
3、ntheimage.ThispaperappliesthetheoryofCBIRtoresearchtheretrievalofstoneimage.Accordingtothefeaturesofthesamestone-image,theimageretrievalsystemadoptscolorfeatureasitsretrievalmethodinthispaper.Sincecolorfeatureisrelativelyrobusttobackgroundcomplicationandindependentofimagesizeandorientati
4、on,itiswidelyusedinimagefeaturerepresentations.Inthispaperthecolorspacemodelandthemethodsofcolorfeatureextractingisintroducedfirstly,andthealgorithmsforcolorfeaturerepresentationofstoneimagesareproposed.TheyselectedtheHSLcolorspaceconsistentwithhuman’Svisualcharacteristics,usedthepercept
5、ionofhuman,theweightsofHSLcolorspacearenon.intervalquantified,andthenthevectorsofcolorfeaturesareconstructed.Secondly,thesimilaritymeasurealgorithmsareproposedindetail.Finally,thealgorithmsproposedinthepaperaretestedandevaluated.Theresultsdemonstratethattheimprovedapproachcanobtainbetter
6、retrievalresultsthanothers.ThisissueisfundedbyFujianProvinceScientificProject“ResearchonKey-technologiesofStoneRetrievalBasedonthesimilarityofimages'’.Thetestingsystemisdevelopedandusedtotesttheretrievaleffectofalgorithms·ThedevelopingsystemisWindowsXPandthedevelopmentenviromentisVisualn
7、武汉理工大学硕士学位论文Studio.NET2003ofMicrosoftCompany.Theresultsoftesthlgal'esatisfied.FurtherworkwillbedonetodevelopthissystemSOastoreleaseanappliedgovernmentinformationretrievalsyStem.K叫word:CBIR,ColorfeatureExtraction,Similaritymeasure,StoneimageI/I独创性声明本人声明,所呈交的论文是我个人在导师指导下进行的
此文档下载收益归作者所有