简介:分布式视频编码具有编码简单、高效压缩等特点,近些年国内外设计实现了多种分布式视频编码算法.本文提出一种基于变换域的分布式视频编码仿真测试软件设计方案,以测试QICF、CIF以4:2:0采样的YUV原始视频文件,通过改变量化参数、提取位平面参数改变视频文件传输比特率,以测试该算法在不同比特率下的率失真性能,通过更改量化表和运动预测搜索半径,搜索块大小,以适用不同序列更改设置,达到最佳性能.该方案可以验证分布式视频编码中各个模块的功能,并可以提出相应模块的各种改进算法验证其性能.
简介:Inthenormaloperationcondition,aconventionalsquare-rootcubatureKalmanfilter(SRCKF)givessufficientlygoodestimationresults.However,ifthemeasurementsarenotreliable,theSRCKFmaygiveinaccurateresultsanddivergesbytime.ThisstudyintroducesanadaptiveSRCKFalgorithmwiththefiltergaincorrectionforthecaseofmeasurementmalfunctions.Byproposingaswitchingcriterion,anoptimalfilterisselectedfromtheadaptiveandconventionalSRCKFaccordingtothemeasurementquality.Asubsystemsoftfaultdetectionalgorithmisbuiltwiththefilterresidual.Utilizingaclearsubsystemfaultcoefficient,thefaultysubsystemisisolatedasaresultofthesystemreconstruction.Inordertoimprovetheperformanceofthemulti-sensorsystem,ahybridfusionalgorithmispresentedbasedontheadaptiveSRCKF.Thestateanderrorcovariancematrixarealsopredictedbythepriorifusionestimates,andareupdatedbythepredictedandestimatedinformationofsubsystems.Theproposedalgorithmswereappliedtothevesseldynamicpositioningsystemsimulation.TheywerecomparedwithnormalSRCKFandlocalestimationweightedfusionalgorithm.ThesimulationresultsshowthatthepresentedadaptiveSRCKFimprovestherobustnessofsubsystemfiltering,andthehybridfusionalgorithmhasthebetterperformance.Thesimulationverifiestheeffectivenessoftheproposedalgorithms.