简介:针对1点RANSAC(RandomSampleConsensus)单目视觉EKF(ExtendedKalmanFilter)算法中的滤波发散问题,分析了滤波发散的产生原因,提出了一种基于渐消记忆滤波的1点RANSAC单目视觉姿态估计算法。该算法通过在EKF滤波方程中引入加权因子,逐渐加大当前数据的权重,相应地减少旧数据的权重,有效地扼制了算法中的滤波发散问题。最后通过两组验证性实验验证说明了算法的有效性。实验结果表明:该算法能够有效地解决1点RANSAC单目视觉EKF算法中的滤波发散问题,具有更高的精度。第一组双轴联动实验,航向角的平均误差减小2.4158?,俯仰角平均误差减小0.1782?;第二组偏航轴大角度转动实验,摄像机航向角的估计误差一直保持在1.5?以内。
简介:TiO2-loadedactivatedcarbonfibers(ACF)werepreparedbyahydrothermalmethod.Thesampleswerecharacterizedbyscanningelectronmicroscopy(SEM),X-raydiffraction(XRD),Fouriertransforminfrared(FT1R)spectrometryandUV-visdiffusereflectancespectra(DRS).SEMimagesshowedthattheTiO2nanoparticlesweredepositedonthesurfaceofACF,andtheparticlesizeandloadingamountofTiO2werevariedbychangingtheinitialconcentrationoftetrabutyltitanate(TBOT).TheresultsofanashexperimentshowedthattheloadingamountsofTiO2were18.4%,43.3%,52.5%,75.1%,and91.1%forinitialconcentrationsofTBOTof0.07,014,0.21,0.28,and0.35mol/L,respectively.PhysicalinteractionsplayedanimportantroleintheformationofTiO2/ACFcompositefibersthatabsorbUVandvisiblelight.ComparedwiththoseofACF,improvedadsorptionandphotocatalyticactivitytowardRhodamineB(RhB)wereobservedforTiO2/ACFcompositefiber.TheRhodamineBcouldberemovedefficientlybyTiO2/ACFcompositefibers,andtheTiO2loadingamounthadasignificanteffectonthephotocatalyticactivityofTiO2/ACFcompositefibers.
简介:Fortwo-dimensionnonlinearconvectiondiffusionequation,atwo-gridmethodofcharacteristicsfinite-elementsolutionwasconstructed.Inthismethodthenonlineariterationsisonlytoexecuteonthecoarsegridandthefine-gridsolutioncanbeobtainedinasinglelinearstep.Forthenonlinearconvection-dominateddiffusionequation,thismethodcannotonlystabilizethenumericaloscillationbutalsoacceleratetheconvergenceandimprovethecomputationalefficiency.Theerroranalysisdemonstratesifthemeshsizesbetweencoarse-gridandfine-gridsatisfythecertainrelationship,thetwo-gridsolutionandthecharacteristicsfinite-elementsolutionhavethesameorderofaccuracy.Thenumericalexampleconfirmsthatthetwo-gridmethodismoreefficientthanthatofcharacteristicsfinite-elementmethod.
简介:Aself-adaptiveprecisealgorithminthetimedomainwasemployedtosolve2-Dnonlinearcoupledheatandmoisturetransferproblems.Byexpandingvariablesatadiscretizedtimeinterval,thevariationsofvariablescanbedescribedmoreprecisely,andanonlinearcoupledinitialandboundaryvalueproblemwasconvertedintoaseriesofrecurrentlinearboundaryvalueproblemswhicharesolvedbyFEtechnique.Inthecomputation,noadditionalassumptionandthenonlineariterationarerequired,andacriterionforself-adaptivecomputationisproposedtomaintainsufficientcomputingaccuracyforthechangesizesoftimesteps.Inthenumericalcomparison,thevariationsofmaterialpropertieswithtemperature,moisturecontent,andbothtemperatureandmoisturecontentaretakenintoaccount,respectively.Satisfactoryresultshavebeenobtained,indicatingthattheproposedapproachiscapableofdealingwithcomplexnonlinearproblems.
简介:Carbonnanotube(CNTs)/Fe-Ni/TiO2nanocompositephotocatalystshavebeensynthesizedbyaninsitufluidizedbedchemicalvapordeposition(FBCVD)method.ThecompositephotocatalystswerecharacterizedbyXRD,Ramanspectroscopy,BET,FESEM,TEM,UV-visspectroscopy,andXPS.TheresultsshowedthattheCNTsweregrowninsituonthesurfaceofTiO2.Fe(lll)inTiO2showednochemicalchangesinthegrowthofCNTs.Ni(Ⅱ)waspartlyreducedtometalNiintheFBCVDprocess,andthemetalNiactedasacatalystforthegrowthofCNTs.ThephotocatalyticactivitiesofCNTs/Fe-Ni/TiO2decreasedwiththeriseoftheFBCVDreactiontemperature.ForthesamplesynthesizedatlowFBCVDtemperature(500℃),morethan90%andnearly50%ofmethylenebluewereremovedunderUVirradiationin180minandundervisiblelightirradiationin300min,respectively.TheprobablemechanismofsynergisticenhancementofphotocatalysisontheCNTs/Fe-Ni/TiO2nanocompositeisproposed.
简介:从模式识别的角度分析了搜索模式下水下运载体的重力匹配问题,利用模式识别神经网络实现重力匹配定位。在重力图匹配时,以惯性导航仪指示位置为中心规划真实位置的网格点搜索范围,从参考重力图上提取相应一系列的重力数据,与对应网格点的位置一起定义成多个模式类,构造相应的模式识别概率神经网络,运用该神经网络将实时重力测量数据识别到某个模式类,对比模式类的定义确定载体位置。在实测重力图上对重力辅助惯性导航系统进行了计算机仿真研究。结果表明,在重力场特征显著区域该重力匹配算法能够有效减小厄特弗斯效应的影响,其导航系统定位误差小于一个重力图网格,匹配率在80%以上,匹配效果优于一般的相关匹配算法。
简介:在不同工况下,旋转爆震波能够以单波、双波、多波模式进行传播.但在同一工况下,是否存在不同模式的稳定传播爆震波还有待进一步研究.基于Euler方程,耦合氢气/空气的有限化学反应速率模型,并采用高分辨率的5阶有限差分格式WENO-PPM5离散对流项,对三维旋转爆震波进行了数值模拟.计算结果表明,在同一特定工况下,旋转爆震波能够以两种不同的传播模式稳定传播,即单波模式和双波模式.详细地对比了两种传播模式下的流场特征、爆震波传播特性、推力性能等.在同一工况下,两种传播模式的爆震波周向传播速度相差不多,但双波模式的频率约为单波模式的2倍;双波模式下质量流量、比冲、推力的平均值均略高于单波模式;且双波模式的可燃混气层高度约为单波模式的1/2,这有助于缩小旋转爆震发动机的长度,使之更加紧凑.