基于H.264标准的运动估计算法研究

《基于H.264标准的运动估计算法研究》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、西南交通大学硕士学位论文基于H.264标准的运动估计算法研究姓名：王立飞申请学位级别：硕士专业：通信与信息系统指导教师：唐小虎20080701西南交通大学硕士研究生学位论文第ll页I

2、——一————————一，——IllllllllII●AbstractThemotionestimationresearchonvideocompression，whichisbasedonthenewH．264videocompressionstandard，hasbeenquiteinterestedbyresearchersinvideocompressionf

3、ield．Theeffectofefficiencyformotionestimationalgorithmmainlyliesinthreeareas，namelyqualityofreconstructedimage，compressionrateandspeedofsearching．Consideringtheaccuracyofsearch，iftheestimationforaccuratemotionismoreaccurate，thenthequalityofthepredictivecompensationimageisbette

4、r，themargincompensationandthebitsforcompensationissmaller,andthebitrateislower．Consideringtherateofsearch，thehighertherateis，thebetterapplicationinthereal—timeenvironmentiS．Therefore．theresearchdirectionformotionestimationalgorithmistodoimprovementsonbothaccuracyandrateforsear

5、ch．ThisessaystudiesonthetechnicalcharacteristicsofH。264standard，andprinciplesofmotionestimationandmotionCompensation．Firstly，ananalysisforthealgorithmsoftheclassicthree·stepsearch，thenewthree—stepsearch，thediamondsearchandthehexagonsearchisgiven．Thenbasedontheimprovedalgorithm

6、stillnow，theauthorpointsouttheresearchdirectionforthemotionestimationalgorithm．BasedonthestudyoftheUMHexagonSalgorithmandaimedtotheshortcomingofexcessivesearchpointsandinsufficiency,thispaperproposessomeimprovement：becausethehelixsearchof25pointsdonotconsidersthespecificmotion

7、vectordistribution，thesearchpointsisrevisedtoninepoints；thenon—-uniformmulti··storeyhexagonsearchdonottakeadvantageoftheimagemotioncharacteristic，whichleadstoexcessivesearchpoints，SOthispaperproposesanewsearchcyclostyle；theunsymmetrical—crosssearchismodifiedtosymmetrical·cross

8、search。酉南交通大学硕士研究生学位论文第lll页IIllllIll——一——一i___●_-_--_--_-_-____-__-●_____--___-●____--Thenewalgorithmmakesuseofvariousoftestsequences，makessurethatthePSBRandbitrateguaranteed，andlowerstimeforsearchandmotionestimation．Thispaperalsoresearchesotherimprovementintheworldandhasbette

9、rresultscomparedwiththetwoimprovementwaysinthem．KeyWords：Moti