基于NAO机器人目标识别与定位算法_柏雪峰

1671-1742201406-0625-05NAO610031NAO、。、360YUVNAO。NAOTP249A2014-09-1701－3NAO、。NAO、。1NAO1．1NAO1NAOv4AldebaranＲoboticsNAOv458cm2514。1．2Nao、32。、。NaoQiNAO。NaoQi、、。Nao2。NaoNaoQi24PCNaoQi34。296201412JOUＲNALOFCHENGDUUNIVEＲSITYOFINFOＲMATIONTECHNOLOGYVol．29No．6Dec．2014231．3NAONAOv4YUV4225。YUV422UV32。NAO4Y0U0V0Y1U1V1Y2U2V2Y3U3V3Y0U0Y1VIY2U2Y3V3Y0U0VIY1U0VIY2U2V3Y3U2V36。22．1NAO1。NAO。2360°。。。4、3。WalkTo010α。。。4。2．22．2．1YUV、37。。、、。ＲGB、HSV、YUV、CMY5。。NAOＲGB“、”。ＲGB－YU、V7。8。2．2．2。、。。62629“”“”。。。YUV422UV25505。。、、、、、、842552550。6。5ＲGB6、2．2．3178NAOv3NAOhαr79。7NAOSS=rsinα2－H－r槡21。2NAOv48。8tanα+β=h－ts2S=h－rtanα+β3hh=459．59mmαHeadPitchβNAOβ=39．74rS。h=459．59mm、β=39．74α3S=0．45959－rtan+39．7°4rS7266NAOS≈0．45959tan+39．7°59NaoS。SS、3。3、。NAOv4NAO47．64°60．9°2。xy640*480320240βα。10－15。αα=x××π1806x=x－3206407=60．9°86－8α=x－320640×60．9°×π1809xyx9NAOααWalkTo010α2β、。ββ=y－240480×47．64°×π18010yxyy10NAOβNAOHeadPitch。2．2．4。“”。、3。。。32120。3。9。NAO。。3826293102。、。1．J．20092612．2．J．20102788－27．3．J．2009291010－29．4AldebaranCommunityWebsiteEB/OL．https//community．aldebaran．com/．5OpenCVEB/OL．http//．opencv．org．cn/．6．YUVJ．2013379．7LeilaSabetiEhsanParviziQMJonathanWu．VisualTrackingUsingColorCamerasandTime-of-FlightＲangeImagingSensorsJ．JournalofMultimedia20083228－36．8．NAOJ．2013827－29．9．J．20082410223－225．10XiangYinNoboruNoguchiJongminChoi．DevelopmentofatargetrecognitionandfollowingsystemforafieldrobotJ．ComputersandElectronicsinAgricultureIF1．7662013Elsevier．11．ＲobocupJ．201046－11．12OlivierMichelYvanBourquin．Jean-ChristopheBaillieＲoboCup2008ＲobotSoccerWorldCupXIIB．LectureNotesinComputerScience2009Springer．13．J．200943111975－1980．14．ＲobocupD．2010．15．ＲobocupNAOJ．20132612．TargetＲecognitionandLocalizationAlgorithmbasedonNAOＲobotBAIXue-fengYANGBinSchoolofInformationScienceandTechnologySouthwestJiaotongUniversityChengdu610031ChinaAbstractAccordingtoanewtypeofhumanoidrobotandthemostwidelyusedNAOplatformanimprovedidentificationmethodisproposed．Itreducestheeffectoflightontargetrecognitionshortensthereactiontimeofrobotidentificationlocalizationandtrackingwhichisawholeclosed-loopoperationprocess．Thesearchprocessusesatrilogyincludingin-si-tusearchwalkingandrotatinglookuptofufilla360°fullvisualtracking．TargetrecognitionisbasedonYUVcolorspace．Targettrackingisimplementedwiththealgorithmthatalwayskeepsthetargetinthecenterofthevisionfield．TheNAOv4HeadPicthjointangleandthedistancecalculationformulaarealsoanalyzedtoachievefastlocalization．KeywordscomputerapplicationtechnologyembeddedNAOrobotclosed-loopoperationtargetrecognitiontargetlocalization9266NAO