kinect windowshardware和sdk概述外文文献翻译学位论文Word格式文档下载.docx
《kinect windowshardware和sdk概述外文文献翻译学位论文Word格式文档下载.docx》由会员分享,可在线阅读,更多相关《kinect windowshardware和sdk概述外文文献翻译学位论文Word格式文档下载.docx(10页珍藏版)》请在冰豆网上搜索。
![kinect windowshardware和sdk概述外文文献翻译学位论文Word格式文档下载.docx](https://file1.bdocx.com/fileroot1/2023-2/1/cc40b9b0-4b6a-42cb-98f4-28e75809f539/cc40b9b0-4b6a-42cb-98f4-28e75809f5391.gif)
functionalities,properties,andlimits
•SoftwarearchitecturedefiningtheKinectSDK1.6MotioncomputingandKinect.BeforegettingKinectinmotion,let'
strytounderstandwhatmotioncomputing(ormotioncontrolcomputing)isandhowKinectbuiltitssuccessinthisarea.Motioncontrolcomputingisthedisciplinethatprocesses,digitalizes,anddetectsthepositionand/orvelocityofpeopleandobjectsinordertointeractwithsoftwaresystems.
MotioncontrolcomputinghasbeenestablishingitselfasoneofthemostrelevanttechniquesfordesigningandimplementingaNaturalUserInterface(NUI).NUIsarehuman-machineinterfacesthatenabletheusertointeractinanaturalwaywithsoftwaresystems.ThegoalsofNUIsaretobenaturalandintuitive.NUIsarebuiltonthefollowingtwomainprinciples:
•TheNUIhastobeimperceptible,thankstoitsintuitivecharacteristics:
(asensorabletocaptureourgestures,amicrophoneabletocaptureourvoice,andatouchscreenabletocaptureourhands'
movements).Alltheseinterfacesareimperceptibletousbecausetheiruseisintuitive.Theinterfaceisnotdistractingusfromthecorefunctionalitiesofoursoftwaresystem.
•TheNUIisbasedonnatureornaturalelements.(theslidegesture,thetouch,thebodymovements,thevoicecommands—alltheseactionsarenaturalandnotdivertingfromournormalbehavior).NUIsarebecomingcrucialforincreasingandenhancingtheuseraccessibilityforsoftwaresolution.ProgrammingaNUIisveryimportantnowadaysanditwillcontinuetoevolveinthefuture.KinectembracestheNUIsprincipleandprovidesapowerfulmultimodalinterfacetotheuser.Wecaninteractwithcomplexsoftwareapplicationsand/orvideogamessimplybyusingourvoiceandournaturalgestures.Kinectcandetectourbodyposition,velocityofourmovements,andourvoicecommands.Itcandetectobjects'
positiontoo.
MicrosoftstartedtodevelopKinectasasecretprojectin2006withintheXboxdivisionasacompetitiveWiikiller.In2008,MicrosoftstartedProjectNatal,namedaftertheMicrosoftGeneralManagerofIncubationAlexKipman'
shometowninBrazil.The
project'
sgoalwastodevelopadeviceincludingdepthrecognition,motiontracking,
facialrecognition,andspeechrecognitionbasedonthevideorecognitiontechnology
developedbyPrimeSense.KinectforXboxwaslaunchedinNovember2010anditslaunchwasindeedasuccess:
itwasanditisstillabreak-throughinthegamingworldanditholdstheGuinnessWorldRecordforbeingthe"
fastestsellingconsumerelectronicsdevice"
aheadoftheiPhoneandtheiPad.InDecember2010,PrimeSense()releasedasetofopensourcedriversandAPIsforKinectthatenabledsoftwaredeveloperstodevelopWindowsapplicationsusingtheKinectsensor.Finally,onJune172011MicrosoftlaunchedtheKinectSDKbeta,whichisasetoflibrariesandAPIsthatenableustodesignanddevelopsoftwareapplicationsonMicrosoftplatformsusingtheKinectsensorasamultimodalinterface.WiththelaunchoftheKinectforWindowsdeviceandtheKinectSDK,motioncontrolcomputingisnowadisciplinethatwecanshapeinourgarages,writingsimpleandpowerfulsoftwareapplicationsourselves.Thisbookiswrittenforallofuswhowanttodevelopmarket-readysoftwareapplicationsusingKinectforWindowsthatcantrackaudioandvideoandcontrolmotionbasedonNUI.InanareawhereKinectestablisheditselfinsuchashortspanoftime,thereistheneedtoconsolidateallthetechnicalresourcesanddeveloptheminanappropriateway:
thisisourzero-to-heroKinectinmotionjourney.Thisiswhatthisbookisabout.
ThisbookassumesthatyouhaveabasicknowledgeofC#andthatweallhaveagreatpassiontolearnaboutprogrammingforKinectdevices.ThisbookcanbeenjoyedbyanybodyinterestedinknowingmoreaboutthedeviceandlearninghowtotrackaudioandvideousingtheKinectforWindowsSoftwareDevelopmentKit(SDK)1.6.Wedeeplybelievethisbookwillhelpyoutomasterhowtoprocessvideodepthandaudiostreamandbuildmarket-readyapplicationsthatcontrolmotion.Thisbookhasdeliberatelybeenkeptsimpleandconcise,whichwillaidyoutoquicklygraspthecoreandcriticalconcepts.
BeforejumpingonthecoreofaudioandvisualtrackingwithKinectforWindows,let'
stakethespaceofthisintroductionchaptertounderstandwhatthehardwareandsoftwarearchitecturesKinectforWindowsanditsSDK1.6use.
Hardwareoverview
TheKinectdeviceisahorizontalbarcomposedofmultiplesensorsconnectedtoabasewithamotorizedpivot.ThefollowingimageprovidesaschematicrepresentationofallthemainKinecthardwarecomponents.LookingattheKinectsensorfromthefront,fromtheoutsideitispossibletoidentifytheInfrared(IR)Projector
(1),theRGBcamera(3),andthedepthcamera
(2).Anarrayoffourmicrophones(6),thethree-axisaccelerometer(5),andthetiltmotor(4)arearrangedinsidetheplasticcase.
ThedeviceisconnectedtoaPCthroughaUSB2.0cable.ItneedsanexternalpowersupplyinordertoworkbecauseUSBportsdon'
tprovideenoughpower.Nowlet'
sjumpintothemainfeaturesofitscomponents.
TheIRprojector
TheIRprojectoristhedevicethatKinectusesforprojectingtheIRraysthatareusedforcomputingthedepthdata.TheIRprojector,whichfromtheoutsidelookslikeacommoncamera,isalaseremitterthatconstantlyprojectsapatternofstructuredIRdotsatawavelengtharoundof830nm(patentUS20100118123,PrimeSenseLtd.).Thislightbeamisinvisibletohumaneyes(thattypicallyrespondtowavelengthsfromabout390nmto750nm)exceptforaredbrightdotinthecenterofemitter.Thepatterniscomposedby3x3subpatternsof211x165dots(foratotalof633x495dots).Ineachsubpattern,onespotismuchbrighterthanalltheothers.
Asthedottedlight(spot)hitsanobject,thepatternbecomesdistorted,andthisdistortionisanalyzedbythedepthcamerainordertoestimatethedistance
betweenthesensorandtheobjectitself.
Depthcamera
Thedepthcameraisa(traditional)monochromeCMOS(complementary
metal-oxide-semiconductor)camerathatisfittedwithanIR-passfilter(whichisblockingthevisiblelight).Thedepthcameraisthedevicethat
Kinectusesforcapturingthedepthdata
Thedepthcameraisthesensorreturningthe3Dcoordinates(x,y,z)ofthesceneasastream.ThesensorcapturesthestructuredlightemittedbytheIRprojectorandthelightreflectedfromtheobjectsinsidethescene.Allthisdataisconvertedintoastreamofframes.EverysingleframeisprocessedbythePrimeSensechipthatproducesanoutputstreamofframes.Theoutputresolutionisupto640x480pixels.Eachpixel,basedon11bits,canrepresent2048levelsofdepth.
TheRGBcamera
TheRGBcameraissimilartoacommoncolorwebcam,butunlikeacommonwebcam,theRGBcamerahasn'
tgotanIR-cutfilter.ThereforeintheRGBcamera,theIRisreachingtheCMOS.Thecameraallowsaresolutionupto1280x960pixelswith12imagespersecondspeed.Wecanreachaframerateof30imagespersecondataresolutionof640x480with8bitsperchannelproducingaBayerfilteroutputwithaRGGBDpattern.Thiscameraisalsoabletoperformcolorflickeravoidance,colorsaturationoperations,andautomaticwhitebalancing.Thisdataisutilizedtoobtainthedetailsofpeopleandobjectsinsidethescene.
Microsoftnow
announced
ithasstarteddeliveringKinectforWindowsv2DeveloperPreviewkitstodeveloperswhosignedup
forthe$399program.TheDeveloperPreviewincludesapre-releaseKinectforWindowsv2sensorandthenextreleaseoftheKinectforWindowssoftwaredevelopmentkit(SDK).
Microsoftsaysitreceived“thousandsofapplications”fortheprogramandselectedanunknownnumber(again,“thousands”)ofparticipantsbasedontheir“expertise,passion,andtherawcreativityoftheirideas.”Thecompanyispromisingongoingupdatesaswellasaccesstoprivateforumswheredeveloperscandiscusswhatworksandwhatdoesn’t.
Thebestpart,however,isthatparticipantswillalsoreceiveaKinectforWindowsv2sensorwhentheybecomeavailable“nextsummer.”Microsoftwouldn’tshareamorespecifictimeframethanthat.
TheKinectforWindowsv2sensorwillfeaturethecorecapabilitiesofthenewKinectforXboxOnesensor,includinganultrawide-angle1080pHDdepthcamera.ThedevicecapturesRGBcolorvideoataround30framespersecondandcanevenmonitortheuser’sheartrate.Hereishowthecompanybreaksdowntheimprovements:
RealVision:
KinectRealVisiontechnologydramaticallyexpandsitsfieldofviewforgreaterlineofsight.Anall-newactiveIRcameraenablesittoseeinthedark.Andbyusingadvancedthree-dimensionalgeometry,itcaneventellifyou’restandingoffbalance.
RealMotion:
KinectRealMotiontechnologytrackseventheslightestgestures.Soasimplesqueezeofyourhandresultsinprecisecontroloveranapplication,whetheryou’restandinguporsittingdown.
RealVoice:
KinectRealVoicetechnologyfocusesonthesoundsthatmatter.Thankstoanall-newmulti-microphonearray,theadvancednoiseisolationcapabilityletsthesensorknowwhotolistento,eveninacrowdedspace.
ItbeatstheoriginalKinectonalmosteverylevel.MicrosoftbelievesitsNaturalUserInterface(NUI)benefitswillimpactmultipleindustries,fromretailandmanufacturingtohealthcare,education,communications,andsoon.Wedon’tdoubt