Observation and analysis of large-scale human

motion

This is a preprint version of the article to Human Movement Science, with preprint formatting. For the official version of the paper please see: J. Per , M. Bon, S. Kovacic, M. ibila, B. De man: "Observation and Analysis of Large-scale Human Motion", Human Movement Science, 21(2), 295-311, July 2002.

Observationandanalysisoflarge-scalehuman

motion

JanezPerˇsa, ,1,2,MartaBonb,2,StanislavKovaˇciˇca,1,2,

b,2ˇMarkoSibilaandBrankoDeˇzmanb,2

aFacultyofElectricalEngineering,UniversityofLjubljana,Trˇzaˇska25,

SI-1000Ljubljana,Slovenia

bFacultyofSport,UniversityofLjubljana,Gortanova22,SI-1000Ljubljana,Slovenia

Abstract

Manyteamsportsincludecomplexhumanmovement,whichcanbeobserved

atdi erentlevelsofdetail.Someaspectsoftheathlete’smotioncanbestudied

indetailusingcommerciallyavailablehigh-speed,high-accuracybiomechani-

calmeasurementsystems.However,duetotheirlimitations,thesedevicesare

notappropriateforstudyinglarge-scalemotionduringagame(forexample,

themotionofaplayerrunningacrosstheentireplaying eld).Wedescribe

analternativeapproachtostudyingsuchlargescalemotion,andpresenta

Correspondingauthor.Tel.:+386-1-4768-876;fax:+386-1-4768-279;email:

janez.pers@fe.uni-lj.si

1SupportedbytheMinistryofScienceandTechnologyoftheRepublicofSlovenia

(Researchprogram1538-517).

2SupportedbytheSlovenianFoundationforSport(Project176/00).

PreprintsubmittedtoElsevierScience4thApril2002

motion

video-based,computer-aidedsystem,developedspeci callyforthepurposeofacquiringlarge-scalemotiondata.Thebaselineofourapproachconsistsofsacri cingmuchofthespatialaccuracyandtemporalresolutionofwidelyusedbiomechanicalmeasurementsystems,toobtaindataonhumanmove-mentthatspanlargeareasandlongintervalsoftime.Datacanbeobtainedforeachoftheobservedathleteswithreasonableamountofoperatorwork.Thesystemwasdevelopedusingtherecordingsofahandballmatch.Several eldtestswereperformedtoassessmeasurementerror,includingcomparisontooneofthewidelyavailablebiomechanicalmeasurementsystems.Withthehelpofthesystempresented,wecouldobtainpositiondataforall14handballplayersona40×20meterlargecourtwithRMSerrorbetterthan0.6me-ter,coveringonehourofaction.Severalresults,obtainedduringthehandballmatchstudyarepresented,inordertohighlighttheimportanceoflarge-scalemotionacquisition.

PsycINFOclassi cation:2330,3720,4120

Keywords:AthleticPerformance;ActivityLevel;Measurement;AutomatedInformationProcessing;Evaluation;

1Introductionandmotivation

Mostsportsincludecomplexmotion,whichcanbestudiedatdi erentlevelsofdetail.Thisisespeciallytrueforteamsports(e.g.soccer,handball,basket-ball).Someaspectsofanathlete’smovementcanbestudiedindetailusingcommerciallyavailablehigh-speed,high-accuracybiomechanicalmeasurement

2

motion

systems.AdetailedcomparisonofthesesystemswasmadebyRichards(1999),andthebasicprinciplesbehindsuchmeasurementswerepresentedbyGruen(1997).Thesesystemsprovidehigh-quality,high-resolutiondataonhumanmovement,however,theyalsoimposeseverelimitations,whichmaketheminappropriateforstudyingmotioncoveringlargeareasforlongintervalsoftime.

Inteamsports,actionisoftenspreadacrossthewholeplaying eld(which,forexampleinEuropeanhandball,measures40×20meters),andmatchescanlastforanhourormore.Mostofthebiomechanicalmeasurementsystemsprovideextremelyaccuratedataonhumanmovement,butcannote ectivelycoversuchalargearea.Thesesystemsprovidehightemporalresolution(100Hzormore),butoverrelativelyshortintervalsoftime,whenthedurationofatypicalmatchisconsidered.Manyofthesesystemsrequiresomekindofmarkerstobeattachedtothebodyoftheathlete,whicharedistractingandnotacceptableduringregularleagueorchampionshipmatches.

Forthepurposeoftacticalmatchanalysis,informationaboutmovementoftheathletes(players)participatinginthematchisneeded.Theaccuracyre-quirementsforsuchdataarefarlowerthanforthepurposeofbiomechanicalanalysis.Tenyearsago,researchersusedself-made,video-basedsystemswithresolutionof1meterforsuccessfulanalysisofasoccermatch(Erdmann,1992),however,noerroranalysiswasperformed.Ontheotherhand,itisdesirablethattheacquireddatacoversboththewholeplaying eldandthewholedu-rationofthematch.Applicationofplayermotionanalysisusingconventionalbiomechanicalmotionanalysissystemswouldthereforebeextremelydi cult,costly,andwouldresultinanunacceptableamountofover-accuratedata.

3

motion

Thedevelopmentanduseofavideo-based,computer-assistedmotionacqui-sitionsystem,whichcandeliverthedataaboutplayerpositionandvelocity,thatconformstotherequirementsofteamsportanalysis,isinthecenterofourresearch.Weobservethemotionoftheplayersontheplaying eldonalargescale,bothtemporallyandspatially;http://www.77cn.com.cnrge-scalemotionrepresentsan”envelope”componentoffull-scalemotion.Forexample,asaplayerrunsacrosstheplaying eld,hiscenterofgravityacceleratesanddeceleratesashemakeseachstep.However,wecanpresentthismotiononalargescale,inwhichcaseitiscomprisedofoneac-celerationatthestartofthemotion,anintervalofnearlyconstantvelocityandthe naldecelerationattheendoftherun.

Thereisastrongmotivationforourresearch.Physicaltraininginteamsportsisextremelyimportantfortopperformanceduringthematches,andtrainingshouldbebasedontheknowledgeofthespeci crequirementsofaparticularsport.Playermotiondatacanrevealmanyaspectsofteamplaythatarenotdirectlyvisible:forexample,itcanhighlightthereasonswhysomeathletesperformbetterthanothers,http://www.77cn.com.cnrge-scalemotiondatacanalsobecomparedtoothervariableswecanmonitorandrecordduringthegame,suchasheartrateorexpertobservationsaboutthecourseoftheplay.Thesevariablesareoftenrecordedatmarkedlylowersamplingrates(0.2Hzforheartrate,forexample)thanmotiondataprovidedbybiomechanicalmotionanalysissystems.

Thebaselineofourapproachconsistsofsacri cingmostofthespatialac-curacyandtemporalresolutionofconventionalmotionanalysissystems,toobtaincoverageoflargear …… 此处隐藏：19118字，全部文档内容请下载后查看。喜欢就下载吧 ……