欢迎来到天天文库
浏览记录
ID:43283824
大小:237.73 KB
页数:13页
时间:2019-09-28
《英文文献1 原稿》由会员上传分享,免费在线阅读,更多相关内容在教育资源-天天文库。
1、Contents1.Theory-2-2.Implementation-4-3.Example-4-4.Variationsandextensions-6-4.1Usingthegradientdirectiontoreducethenumberofvotes-6-4.2Kernel-basedHoughtransform-7-4.3Houghtransformofcurves,andGeneralisedHoughtransform-7-4.4Detectionof3Dobjects(Planesandcylinders)-7-4.5U
2、singweightedfeatures-8-4.6Carefullychosenparameterspace-8-4.6.1ImplementationofanEfficientEllipseDetectionAlgorithm-9-5.Limitations-9-6.Seealso-10-References-11-Accessories:-12--12-Houghtransform(FromWikipedia,thefreeencyclopedia)TheHoughtransformisafeatureextractiontechn
3、iqueusedinimageanalysis,computervision,anddigitalimageprocessing.[1]Thepurposeofthetechniqueistofindimperfectinstancesofobjectswithinacertainclassofshapesbyavotingprocedure.Thisvotingprocedureiscarriedoutinaparameterspace,fromwhichobjectcandidatesareobtainedaslocalmaximai
4、naso-calledaccumulatorspacethatisexplicitlyconstructedbythealgorithmforcomputingtheHoughtransform.TheclassicalHoughtransformwasconcernedwiththeidentificationoflinesintheimage,butlatertheHoughtransformhasbeenextendedtoidentifyingpositionsofarbitraryshapes,mostcommonlycircl
5、esorellipses.TheHoughtransformasitisuniversallyusedtodaywasinventedbyRichardDudaandPeterHartin1972,whocalledita"generalizedHoughtransform"[2]aftertherelated1962patentofPaulHough.[3]ThetransformwaspopularizedinthecomputervisioncommunitybyDanaH.Ballardthrougha1981journalart
6、icletitled"GeneralizingtheHoughtransformtodetectarbitraryshapes".1.TheoryInautomatedanalysisofdigitalimages,asubproblemoftenarisesofdetectingsimpleshapes,suchasstraightlines,circlesorellipses.Inmanycasesanedgedetectorcanbeusedasapre-processingstagetoobtainimagepointsorima
7、gepixelsthatareonthedesiredcurveintheimagespace.Duetoimperfectionsineithertheimagedataortheedgedetector,however,theremaybemissingpointsorpixelsonthedesiredcurvesaswellasspatialdeviationsbetweentheidealline/circle/ellipseandthenoisyedgepointsastheyareobtainedfromtheedgedet
8、ector.Forthesereasons,itisoftennon-trivialtogrouptheextractededgefeaturestoanappropriatesetoflin
此文档下载收益归作者所有