资源描述:
《Nyquist–Shannon sampling theorem(奈奎斯特-香农采样定理)》由会员上传分享,免费在线阅读,更多相关内容在应用文档-天天文库。
1、Nyquist–ShannonsamplingtheoremX(ƒ)1IntroductionSamplingistheprocessofconvertingasignal(forex-ample,afunctionofcontinuoustimeand/orspace)intoanumericsequence(afunctionofdiscretetimeand/orspace).Shannon’sversionofthetheoremstates:[1]Ifafunctionx(t)containsnofrequencieshighe
2、rthanB−BBƒhertz,itiscompletelydeterminedbygivingitsordinatesataseriesofpointsspaced1/(2B)secondsapart.Fig.1:MagnitudeoftheFouriertransformofabandlimitedAsufficientsample-rateistherefore2Bsamples/second,functionoranythinglarger.Equivalently,foragivensampleratef⛿,perfectrecon
3、structionisguaranteedpossibleforabandlimitB≤f⛿/2.Inthefieldofdigitalsignalprocessing,thesamplingthe-Whenthebandlimitistoohigh(orthereisnoban-oremisafundamentalbridgebetweencontinuous-timedlimit),thereconstructionexhibitsimperfectionsknownsignals(oftencalled“analogsignals”)
4、anddiscrete-timeasaliasing.Modernstatementsofthetheoremaresome-signals(oftencalled“digitalsignals”).Itestablishesatimescarefultoexplicitlystatethatx(t)mustcontainsufficientconditionbetweenasignal’sbandwidthandthenosinusoidalcomponentatexactlyfrequencyB,orthatsampleratethatp
5、ermitsadiscretesequenceofsamplesBmustbestrictlylessthan½thesamplerate.Thetocapturealltheinformationfromthecontinuous-timetwothresholds,2Bandf⛿/2arerespectivelycalledthesignal.NyquistrateandNyquistfrequency.Andrespectively,Strictlyspeaking,thetheoremonlyappliestoaclassofth
6、eyareattributesofx(t)andofthesamplingequipment.mathematicalfunctionshavingaFouriertransformthatisTheconditiondescribedbytheseinequalitiesiscalledzerooutsideofafiniteregionoffrequencies(seeFig1).theNyquistcriterion,orsometimestheRaabecondi-Intuitivelyweexpectthatwhenoneredu
7、cesacontinuoustion.Thetheoremisalsoapplicabletofunctionsofotherfunctiontoadiscretesequenceandinterpolatesbacktodomains,suchasspace,inthecaseofadigitizedimage.acontinuousfunction,thefidelityoftheresultdependsTheonlychange,inthecaseofotherdomains,istheunitsonthedensity(orsam
8、plerate)oftheoriginalsamples.ofmeasureappliedtot,f⛿,andB.Thesamplingtheoremintroducestheconcepto