简介:聚类的地震指在时空与对方一起仔细发生的地震事件。因为他们的重叠波形记录使拣第一到达困难,聚类地震的hypocenters不能被traveltime地点方法精确地决定。这里,我们使用一个反向时间的成像(RTI)方法印射聚类地震。拿直接使用波形的优点,RTI方法是有能力的印射单个小地震或浓密地散布的一些在2-D下面聚类地震地震数组。在3-D情况中,RTI方法成功地被使用从一套稀疏地分布式的表面车站用数据定位长偏移量的马甲地震。为一样的获得几何学,然而,在印射的RTI相遇挑战浓密地散布了聚类地震。当改进印射聚类地震与照明角度的更宽的范围要求一个更稠密的接收装置网络,是明显的时,与合成数据验证RTI方法的实际分辨率是必要的。在我们在三个峡区域的学习区域,我们3-D情况的测试建议在线性排列亚数组下面的一些事件有合理分辨率。
简介:Anewfuzzysupportvectormachinealgorithmwithdualmembershipvaluesbasedonspectralclusteringmethodisproposedtoovercometheshortcomingofthenormalsupportvectormachinealgorithm,whichdividesthetrainingdatasetsintotwoabsolutelyexclusiveclassesinthebinaryclassification,ignoringthepossibilityof'overlapping'regionbetweenthetwotrainingclasses.Theproposedmethodhandlessample'overlap'efficientlywithspectralclustering,overcomingthedisadvantagesofover-fittingwell,andimprovingthedataminingefficiencygreatly.Simulationprovidesclearevidencestothenewmethod.
简介:Renderingoflarge-scaleforestscenesisachallengingtask,whosehighlygeometriccomplexitywillputheavyburdenoncurrentgraphicshardware.Whennavigatingthescene,theoverallvisualresultisgenerallyconsideredasthecoreconcern.Anewmethodisproposedinthispaperforlarge-scaleforestrenderingusingclusteringandmergingstrategies.Ourmethodimprovestherenderingeffectbyclusteringpolygonsaccordingtothepointinformationwithrelationtoneighbours.Afastforestrenderingsystemisdevelopedaccordingly.Therelativetechniquesinthesystemcanimprovethevisualqualityondemandofdifferentapplications.
简介:Theeffectofresinstructureondesalinationofligninsolutionwasinvestigated,theoptimalstructureofresinisasfollows;crosslinkingdegreeas4%,ratioofcationogentoanionogenisnear1.withsuchresinthedesalinationofligninwasproducedverywellbecausetheresinhasbothmoleculesievingandionretardationproperties.Thesulfonationdegressofligninandtotalsaltcontentofligninsolutionweredeterminredwithionexchangetechnique,therelativeerrorlessthan1%.Thesaltcontentofsmallmoleculeintheligninsolutionwascalculatedfromsulfonationdegreeofligninandtotalsalt.Amonggelandmacroporousresinsthebestseparationofligninfromreducingsugarwasachievedwithinterpenetratingsulfonatedresin2×1.5×1.Theseparationofligninwithinterpenetratingresinwascarriedoutsimultaneouslywithfractionationoflignin,theeffectoffractionationwithmacroporoussulfonatedresinisbetterthanthatwithinterpenetratingresin,buttheformerhasadefinitesorptionofligninwhichdecreasedtherecoveryoflignin.
简介:Dempster-Shaferevidencetheory(DStheory)iswidelyusedinbrainmagneticresonanceimaging(MRI)segmentation,duetoitsefficientcombinationoftheevidencefromdifferentsources.Inthispaper,animprovedMRIsegmentationmethod,whichisbasedonfuzzyc-means(FCM)andDStheory,isproposed.Firstly,theaveragefusionmethodisusedtoreducetheuncertaintyandtheconflictinformationinthepictures.Then,theneighborhoodinformationandthedifferentinfluencesofspatiallocationofneighborhoodpixelsaretakenintoconsiderationtohandlethespatialinformation.Finally,thesegmentationandthesensordatafusionareachievedbyusingtheDStheory.ThesimulatedimagesandtheMRIimagesillustratethatourproposedmethodismoreeffectiveinimagesegmentation.
简介:Invideoinformationretrieval,keyframeextractionhasbeenrecognizedasoneoftheimportantresearchissues.Althoughmuchprogresshasbeenmade,theexistingapproachesareeithercomputationallyexpensiveorineffectiveincapturingsalientvisualcontent.Inthispaper,wefirstdiscusstheimportanceofkeyframeextractionandthenbrieflyreviewandevaluatetheexistingapproaches.Toovercometheshortcomingeoftheexistingapproaches,weintroduceanewalgorithmforkeyframeextractionbasedonunsupervisedclustring.Meanwhile,weprovideafeedbackchaintoadjustthegranularityoftheextractionresult.Theproposedalgorithmisbothcomputationallysimpleandabletocapturethevisualcontent.Theefficiencyandeffectivenessarevalidatedbylargeamountofreal-worldvideos.
简介:有在兴趣的一个区域的类似的特征的不同seismo构造的事件的鉴定和分类是在地震危险研究的最重要的题目之一。在这研究,线性、非线性的判别式分析被使用了在伊斯坦布尔的附近分类地震事件。数字速度震动图的垂直部件与发生在2001和2004之间的在1.8和3.0之间的大小(Md)被用于地震事件。二,时间依赖者参数,复杂性和S/P山峰振幅比率作为predictands被选择。线性,二次,diaglinear和diagquadratic判别式功能被调查。方法与的精确性一另外的调整二次的模型是96.6%,96.6%,95.5%,96.6%,和97.6%,分别地与为每个班的各种各样的分类的率。模型的表演与生气确认和替换错误被认为正当。尽管所有模型很好显著地表演了,调整二次的功能从179与就4个分类事件完成了最好的成功率,与复杂方法例如相比甚至更好,自我组织方法,k工具,在文学适用于一样的数据集的Gaussion混合模型。
简介:Indownlinkcoordinatedmulti-point(CoMP)system,fullcooperationisalwaysnotapplicableinrealworldbecauseofitshighrequestinthebackhaul.Todealwiththisproblem,clusteringdecisionismadetoprocesstransmission.Inthispaperclusteringmethodsbasedonthemetricsignal-to-leakage-plus-noise(SLNR)isproposed.Inaddition,userschedulingschemesbasedonSLNRisalsoputuptomaketheschedulingsetaslargeaspossible.Simulationresultsshowthattheproposedclusteringmethodsnotonlyreducethedatasharingamongthecooperatingbasestations(BSs),butalsoimprovethesystemthroughputcomparedwiththetraditionalclusteringmethodsbasedonchannelstrength.
简介:Clustering,asapowerfuldataminingtechniquefordiscoveringinterestingdatadistributionsandpatternsintheunderlyingdatabase,isusedinmanyfields,suchasstatisticaldataanalysis,patternrecognition,imageprocessing,andotherbusinessapplications.Density-basedSpatialClusteringofApplicationswithNoise(DBSCAN)(Esteretal.,1996)isagoodperformanceclusteringmethodfordealingwithspatialdataalthoughitleavesmanyproblemstobesolved.Forexample,DBSCANrequiresanecessaryuser-specifiedthresholdwhileitscomputationisextremelytime-consumingbycurrentmethodsuchasOPTICS,etc.(Ankerstetal.,1999),andtheperformanceofDBSCANunderdifferentnormshasyettobeexamined.Inthispaper,wefirstdevelopedamethodbasedonstatisticalinformationofdistancespaceindatabasetodeterminethenecessarythreshold.ThenourexaminationoftheDBSCANperformanceunderdifferentnormsshowedthattherewasdeterminablerelationbetweenthem.Finally,weusedtwoartificialdatabasestoverifytheeffectivenessandefficiencyoftheproposedmethods.
简介:与小铁磁性或顺磁的粒子的低集中清洗气体是为常规过滤的一项困难的任务。一个新其他的过程,磁性的过滤,在这个工作被使用。氧化铁喷雾器被氧化铁粒子的elutriation从由Geldart-C氧化铁粉末和大球形的Geldart-B沙粒子的混合物组成的一张使流体化的床产生。喷雾器借助于由一,二或三个熨斗组成了的一个磁性的过滤器被过滤刺激蹒跚到对方。试验性的安装也包含了一个isokinetic采样系统和一个MicrotracSRA150粒子分析器。为过滤效率的理论表情从考虑对氧化铁粒子起作用的不同力量的一个以前的模型被推出。试验性的过滤效率匹配相当好计算了理论效率。粒子的增加缩放,这被发现,在数字刺激或在更高生产的应用磁场过滤效率直到在一些情况中的100%。在通过磁性的过滤器的所有过滤实验压力落下是很小的。