欢迎来到天天文库
浏览记录
ID:36444266
大小:3.78 MB
页数:91页
时间:2019-05-10
《数字视频技术在智能运输系统中的应用研究》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、⑨秉霸太警博士学位论文数字视频技术在智能运输系统中的应用研究研究生:何铗军导师:黄卫教授专业:道路与铁道工程东南大学交通学院二零零四年五月摘要AbstractWimthedevelopmentofnetworktechnologyandvideocompressionalgorithm.itbecomesanew仃endtOcompressvideointodigitalflowandWansmitthrowIghnetworks.Butthemainstreamproductsofmarketscall
2、notsarisfyneedofATMSbecauseofitsbiglatency,poorvideoqualityandsimplefunction.Thecontentofthisdissertationcanbesummarizedasfollows:1.RealizedaMPEG-2networkCODECsystem.AnembeddedplatformisconstructedusingMCF5307CPUasitscore.Assembletheplatformwithvariousint
3、erfacechips.aMPEG-2networkCODECsystemisrealized.ThedetaileddesignofcircuitsandprogrammingtechniquesofdriversforucLinuxOSisprovidedinthedissertation.2.BasedontheanalysisofdataformatofMPEG-2stream,adissertationdiscussedthefluctuationsinnetworktransmissionan
4、dthetemporaldelayinheritedindigitalvideothoroughly,andbroughtforwardcolTespondingsolutions,whichensuretheplaybacksmoothandlow-delayed,andmakethissystemprecedingthepeers.3.AsthefrequencyresponsepropeaiesoftheCODECsystemprototypecannotsatisfytherequiremento
5、frelatedstandards,twohigh·frequencycompensationalgorithmsbaSedseparatelyonspatialdomainandfrequencydomainareproposed.Inparticular,therealizationofthecompensationalgorithmbasedonspatialdomainisprovided.111eproposedalgorithmsensurethatallthefunctionalperfor
6、mancesoftheCODECsystemareuptograde.4.Theunreliabilityofnetworktransmissioncancauselossofdata.Errorconcealmentisawaytosolvethisproblem.Basedonthecharacteristicsofimagesforhighwaysurveillanceandcontrol,thecorrelationbetweenBframeandtwoneighboringreferencefr
7、amesisusedtoobtainthemotion-vectorvaluesofthelostpart,thenthelostpartisrecoveredwiththecontentofpreviousreferencefranle.ThismechanismCanachievesatisfactoryresults.5.DiscussedthetechnologyofobtainingtrafficdatafromcompressedvideostreaminⅣ田EG-2format.TheDCT
8、coefficientsofgrayscaleandchroma,andthemotionvectorsofmacroblocksareusedtoinspecttheexistenceofvehiclesandmeasuretheirspeed.Comparedtotraditionalvideoinspectors,thisnewmethodhaShighprecisionandstronganti-interferenc
此文档下载收益归作者所有