2 video analysis

8/13/2019 2 Video Analysis

1/25

CHAPTER

MOTIONANALYSISSINGVIDEOCorlJ. Poyfon

INTRODUCTIONFor many decades,cinematography was the most popular measurement tech-"iq". r", those involved in tf,e analysisof human

motion. cine cameras havetraditionally been considered superior to video cameras becauseof their muchg*"i., pi.,lre resoluti,onand hiiher frame rates.However, over the last decade,considerable advanceshave beelnmade in video technology which now makevideo an attractive alternative to cine. Modern video cameras are now able todeliver excellent pl.t.rr. quality (although still not quite as good as cine) andhigh-rp."d -od.is .u.r u.ii.u. frames rates at least comparable to high-speed.ii. .ur.r.r"s. Unlike cine film, most video recording involves -no processingtime and the recorded magesare available for immediate

playback and analysis'Vid.o,up. ,afeveryinexp"ensivewhencomparedtothehighcostofpurchasingu.rJ p.o..rring of .i.t. fim. fhe significant-improvements.maden video camerai..hrrology, clupled with a substlantial all in price _of he hardware over theolr, a..ii., h", l.d to cine camerasbecoming virtually redundant in sport andixercise biomechanics.video recordings of sport and exercise activities are usually made bybiomechanist, i., ori.r to ,r.,dertake a detailed analysis of an individual'smovement patterns' Although on-line systems Chapter 3) provide an attractivealternative to video' "s aLtthod of capturing motion-data' video motion"rrulyri, has a number of practical advantagesover on-line

motion analysisincluding:o Low cost - video analysissystemsare generally considerably cheaper hanon-line systems.r Minimal interference o the performer - video analysiscan be conducted

without the need for any disiurbance to the performer' e'8' attachment ofreflectivemarkers.

Flexon-liundeAllowpermOn-l

Given theable futurVideo anin natureobservation a TV and framdisplayedThe purpmovemenIt may althat needQuavideo recapproachqualitativprocessininvolvesor moreTypical ljoint cenof foot),ordinatesbefore beAdditionacomputinHoweverthe approThe kineperformacan thenperforme(e.g. to eIn techniqucommonmethodjoint reathroughcomputatangular


2/25

MOTION NALYSISSING IDEO 9. Flexibility - video analysis can be used in environments where some

on_line systems would te unable to operate effectively, e.g. outdoors,underwater, in comPetition.o Allows visual feediack to the performer - video cameras provide apermanenr record of the movement that can be viewed immediately.bn-line systemsdo not generally record the image of the performer'

Typical landmarks selected for digitisation are those assumed to representioint...rtr., of rotation (e.g. knee joint centrei, segmentalendpoints (e.g.endoi foot), or external obj..ts (e.g. a sports implement)' Two-dimensional co-ordinates esulting frorn the digitising process are then scaled and smoothedbeforebeing or.d to calculate inear and angular displacement-time histories'Additional ine-atic information (velocities

and accelerations) s obtained by

Given the advantages isted above, video analysiswill remain, for the foresee-able future, an important method of analysing technique in sport and exercise.Vid.o "rr"iysis oi a person's technique may be qualitative or quantitativein nature. Qualitative analysis involves a detailed, systematic and structuredobservation f th. p.rformer's movement pattern. The video image is displayedon a TV monitor or computer screenand observed n real-time' slow motionand frame-by-frame. Often' multiple images, e'g' front and side views' aredisplayedsimultaneously to allow " -or. complete analysis o be undertaken.Thi purpose of this typ. of analysis is often to establish the quality of the,nou.In.rr, being obre.rred n order to provide some feedbackto the performer'It may also be Jsed a, a means of identifying the key performance parametersthat need to be quantified and monitored in future analyses'

euantitativi analysis involves taking detailed measurements from thevideo recording to .nubl. key perform".r.. p"."-etefs to be quantified. Thisapproach ,.qrrir., -or. ,opl,isticated--hardware and software than for aqoutit"tlu. arr"lysis and it is iital to follow the correct data capture and datapro..rrirg procedures.Quantitative_analysiscan be time-consuming as it ofteninuolu* ir"ro"lly digitising a number of body landmarks (typically eighteenor more points fo, u f,rll body model) over a latge number of .video images.

uting the first and second time derivatives of these displacement data.,u.r,"rn. accuracyof thesederivativeswill be severelycompromised unless

appropriat. data processing echniques are used (discussed n Chapter 7) 'kinematic information obiained from video can be used to quantify keyrmanceparameters(e.g.a take-off angle during a jump)' Such.parametersthen be'compared be"tween performers (e.g. novice vs. elite), within,*.r, 1..g. aiig.r.d.,rs.,torr-faiigued),r monitoredovera periodof timeto evaluatehe effects f trainingover a season)'In order to understand the underlying causesof a given sport or exerciseique,more detailed quantitative analysesare often undertaken' tl: *: *upprou.h s thai of inversedynamics discussedn Chapter9)' Thisinuolu.t computing kinetic information on the performer (e'g' netreactionforces and .r.', -oln.n,s) from kinematic information obtainedeh uid.o, or some other form of motion analysis.The inverse.dynamics,it"tiorr"i procedures require second time-derivative data, i'e' linear andlar accelerations, or the body segmentsbeing analysed' and also require


3/25

1O CARLJ. AYTONvalid body segment inertia data (e'g' mass and moment of inertia)' Thecalculated joint momen-tsurra ror.., Ju., b. subiect to significanterrors unless;;;; ;;;; i, t"k.r, to. minimise the error in the kinematic and inertia data'The interpre,u,ron oi the results of an inverse dynamics analysis is notas straightforward "r"i;;; kin"-uti. analysis. Inverse dynamics

provides aninsight into the n., "ri.., of a[ the ..,urcre, crossing a ioint,

but it does notalrow the .o,,'po,",io.r"-;it;". ;.ntact forces o.1h. torque produced byindividual muscles,or muscle groups' around the joint' Although there are anumber of limitationr-," irr. i".[.re'dy.r"-i.s approach (e.g. winter, 1990), themethod can still provide the biomeciranist wiih a much better understandingof the musculo-skeletal forces and torques acting during a sport or exercise;;;;;y;;h"n could be obtained from an analysis of the movement patternsalone.

Picture quaA video imA full videup of thethe even-nmethods:techniquescan cameframe areformat. I7displayingup a videoextent, deThercan sometcamera puplayer. Th(exceptFraCouleur AEuropeanof pixels.StandardsJapan, andis thereforthat the vethan thesePicture quarefers to thIn theon the domformat alloback on DVaffordableresolutionis importan7201/1080ifor exampledescribed,have varyin

VHS,lines.S-VHDigitaHigh Deither

EQUI MENT ONSI ERATIONSSelectionof the appropriate equipment is important when undertaking a motionanalysis st,rdy ,rring uid.o. The key components of a video motion analysissystemare:. Video camera - to capture images of the movement;r Recording and ,ro."g. device

-- to record and store the images from-the camera. This mJy be an integral part of the video camera itself(camcorder)or an external unit, e'g' hard-disc;. pluyb"ck system- to allow the video images o be viewed for qualitativeor quantitative analYsis;. co-ordinate digitisei - to allow measurements o be taken from the videoimages;. p.oJ.srirrg ".rd analysissoftware

- to enable the user to quantify selectedparametersof the movement.

Videocomeroswhen selecting a video camera with the intention of undertaking a biome-chanical analy"sisof a sport or exercise activity' the important features toconsiderare:

picture qualityframe rate (samPling frequencY)manual high-sPeed huttermanual aPertureadiustmentlight sensitivitygen-lock capabilitYiecording medium (e.g. ape, hard drive)'

aaaaaaa

aaa


4/25

MOTION NALYSISSINGVIDEO 1Picture qualityA video image is made up of a two dimensional array of dots called pixels.A full video image or frame consists of two halves or fields. One field is madeup of the odd-numbered horizontal lines of pixels, the other is made up ofthe even-numbered lines. Video cameras capture an image using one of twomethods: interlaced scan or progressive scan. Cameras that use the interlacetechnique ecord one field first, followed by the second,and so on. A progressivescan camera records a complete frame and the two fields that comprise thisframe are identical. Somecameras have the facility to capture images n eitherformat. 'With progressivescan, the option to analyse a movement at 50H2, bydisplaying ndividual video fields, is lost. The number and size of pixels makingup a video image determine the resolution of the picture and this, to a largeextent, determinesthe picture quality.There are a number of different world standards or video equipment; thiscan sometimes ead to problems of compatibility. For example, a digital videocamerapurchased in the USA, may not be compatible with a UK sourced DVplayer.The phase alternating line (PAL) standard is used in 'S7esternEurope(except rance),Australia and much of East Africa, India and China. SequentialCouleur Avec M6moire (SECAM) is the standard found in France and EasternEuropeancountries. Both PAL and SECAM video have 625 horizontal linesof pixels. This is referred to as the vertical resolution. National TelevisionStandardsCommittee (NTSC) is the standard adopted in North America andJapan,and has 525 lines. The maximum vertical resolution of a video imageis therefore essentially imited by the video standard used. It should be noted

lthat the vertical resolution of a displayed image might be considerably lowerthan hese igures, depending on the specificationof the video equipment used.quality is also influenced by the horizontal resolution of the video. Thisto the number of pixels per horizontal line.In the past couple of years a new video format called HDV has emerged

the domesticmarket and is likely to supercedeexisting st andards. The HDVt allows high definition (HD) video images to be recorded and playedon DV tape. HDV video cameras are now commercially available at very

ble prices and the images produced by these cameras have a verticalion of either 720 or 1080 lines. When purchasing an HDV camera, itimportant to check what mode(s) it can record and playback in (interlaced:1080ior progressive: 20p11.080p)o ensure hat it is compatiblewith,example, he display device. $Tithin each of the world video standards ust;ribed, there are a number of video recording formats available and thesevarying esolutions:

VHS,VHS-Cand 8mm formatseachdeliveraround240-:260horizontallines.S-VHS, -VHS-Cand Hi-B videoprovidearound400 horizontal ines.Digital8 and miniDV deliverat least500 horizontal ines.HighDefinition HD) videogiveseither720 or 1080horizontal ines witheither 280or 1.920 ixelsper line).


5/25

12 CARL.PAYTON

(a) (b)Figure.1 (o)High-speedideo omeroPholronostcomltimo PX) opoble f frome otes p o2OOOHzt ull esolution1024 1024 ixels);b)Comero rocessorni t

Somespecialistvideo cameras (e.g.Photron FastcamUltima APX in Figure 2.1)can record images with resolutionshigher than those describedabove. It shouldbe noted that evenwithin a given recording format, e.g.miniDV, the quality ofthe video image can vary considerably. The resolution of the camera is largelyinfluenced by the quality of its image sensor - the component that convertsthe light from the object into an electrical signal. The most common type ofimage sensor s the charge-coupleddevice(CCD). Most domesticvideo camerashave a single CCD chip, but some higher quality models have three CCDs (onefor each of the primary colours), which result in an improved picture quality.An alternative to the CCD is the complimentary metal oxide semiconductor(CMOS) image sensor.This sensor requires far lesspower than a CCD and isnow used in some standard and high-speedvideo cameras.The specificationof the camera lens s an important factor in determiningpicture quality. Digital video cameras will have both an optical zoom range,i.g.20x and a digital zoom range' e.g.400x. It is important to note thatonce a camera is zoomed in beyond the range of its optical system, he picturequality will drastically reduce and will be unsuitable for quantitative analysis.A...rro.y telephoto lensescan be used to increase he optical zoom of a digitalvideo camera and avoid this problem. They also allows the user to increase hecamefa-to-subjectdistance,whilst maintaining the desired mage size.This willreduce the perspective error although it should be noted that the addition ofa telephoto lens will reduce the amount of light reaching the camera's imagesensor. t is important to check how well a telephoto lens performs at the limitsof the optical zoom, as this is where image distortion will be most pronounced.\X/ide-angleensescan be fitted to video cameras o increase he field of view fora given camera-subjectdistance.However, such lenses end to produce consid-erable image distortion and have limited applications in quantitative analyses.

Frome rote (sompling frequency)In video captufe, the term 'frame' refefs to a complete image captured at aninstant in time (Greaves, L995). Thus the frame rate of a video camera refers

to the numas the sama frame raIf the camwill be comwith a vertone field cappropriateseparatelyaa second n

For socameraswilspeed videoor digital (frame rates100-500H2applications(e.g.Peak Pto RAM (ea computecamera).

One ois the limitewith a storat 2000 Hzseconds.

High speedFor most bspeedshuttthe amountlight. Modeor deactivasampled. 7is exposedperiod of tiThe extentanalysed.It is iThis allowsas it repreactivity thaCollection Poffer shuttenoted that models tha


6/25

MOTION NALYSISSING IDEO ] 3to the number of full images it cap tures per second (this is often referred toas the sampling frequency of the camera). Standard PAL video cameras havea frame rate of 25H2, whereas NTSC cameras have a frame rate of 30H2.If the camera captures using the interlaced scan method, each video framewill be comprised of two video fields (an A and B field). For a video imagewith a vertical resolution of 480 lines, each field would consist of 240 lines,one field comprised of the odd lines, the other of the even lines. With theappropriate hardware or sofrware, it is possible to display the video fieldsseparatelyand sequentially thus enabling measurements o be taken at 1/50 ofa second ncrements (or 1,/60of a second or NTSC), but at reduced resolution.

For some sport and exerciseactivities, he frame rate of conventional videocameraswill be too low and a high-speedvideo camera may be required. High-speedvideo cameras , as with conventional video cameras, can be analogueor digital (seeGreaves,1995 for more detail). Although video cameraswithframe rates beyond 2000H2 are commercially available, cameras with rates of100-500 Hz are generally adequate for most sport and exercise biomechanicsapplications . Although some early high-speed video camerasrecorded to tape(e.g.Peak Performance HSC 200 PS),most models now e ither record theimagesto RAM (e.g. Photron Fastcam Ultima APX shown in Figure 2.1) or direct toa computer hard drive via a Firewire (IEEE) port (e.9. Basler 602f 100H2camera).

One of the major limitations of high-speed cameras hat record to RAM, is he limited recording time available.For example, a high-speedvideo camerawith a storage capacity of B Gb, recording with a resolution of 1,024 x 1024

at 2000 Hz, provides a maximum recording duration of approximately threeseconds.

Highspeed hutterFormost biomechanical pplications, video cameraequippedwith a high-speedhutters essential. he shutter s the component f a camera hat controls

amount of time the camera's mage sensor(e.g.CCD, CMOS) is exposed ot. Modern video camerasuse electronic shuttering,which involves activatingdeactivating he image sensor or a specified ime period, as each video field is. lfhen recording movement using a low shutter speed, heimage sensorexposed o the light passing through the camera lens for a relatively longiod of time; this can result in a blurred or streaked image being recorded.extentof the blurringwould dependon the speed f the movementbeinglysed.It is important that a video camerahas a manual shutter speedoption.is allows the user to select a 'shutter speed' (this term is a misnomerit represents the time the shutter is open) that is appropriate for theivity that is being analysed,and the prevalent lighting conditions (seeData

ion Procedures section of this chapter). Typically, a video camera willshutter speeds anging from tl60-1,/4000 of a second. t should bethat not all video camerasoffer a manual shutter function. Camerathat incorporate a Sports Mode function should be avoided because


7/25

14 CARLJ.AYTONthe shutter speed associatedwith this is often inadequate or fast-movingactivities.

Monuol iris and low-light sensitivityThe iris is the element of the camera's lens systemthat controls the aperture(the adjustable gap in the iris) in order to regulate he amount of light falling oniir. i-is. r.rrrJr.'If too mucL light is permitted to pass hroughthe lens (large"p.r,"rJl, for too long, the result-will be "n orrerexposedmage.

f too little lightf'urr", through th. l"i, (small aperture), the imagewill be underexposed.Videoi"-.ru, genlrally have automatic aperture control that continually adjusts toensure the image is correctly exposed. Some camera models have a manualoverride that aliows the user to specify the aperture setting.This is sometimes.r...rr"ry when conducting biomechanical analyses.For example' when a highshutter speed setting i, .r. d.d in low light conditions' the iris aperture wouldfr"u. ,o b. op.n.d i"id". thun it would be in automatic mode. The drawbackoi aoi"g this is the increasednoise level in the image, which results in a more'grainy'

picture.video cameraseachhave a minimum light level that they require in orderto produce an image. This level is expressed n lux. A camera with a minimumillumination ,r"1,-r..-of lux will perform better in low light conditions than onew i t h a 3 l u x r a t i n g .

Genlock copobilityFor three-dimensional video analysis, it is desirable for the activation of theshutters of the two (or more) cameras to be perfectly synchronised, that is,for the cameras to be gen-locked.This involves physically linking the cameras*irt " gen-lock cable."Unfortunately,most standard video camcorders

do notfr"r. ifr". facility to be gen-locked, although some more expensive models dooffer this feature 1e.g.banon XL H1 HDV 1080i camera). If video camerascannot be gen-lock.i, th. two-dimensional co-ordinates obtained from eachof the camera views must be synchronisedby interpolating the data and thenshifting one data set by the time lag betweenthe camera shutters.The time lagwifl bJ no more than half the reciprocal of frame rate of the camera (e.g. at)sn ,the time lag will be


8/25

MOTION NALYSISSING IDEO 15RecordingmediumImages rom video camerashave traditionally beenrecordedonto some form oft"p., fot example, S-VHS and miniDV. In recent years a number of alternativereiording fot-ats have emerged.Video cameras hat record straight to a smallDVD are more geared oward the home movie-maker, han those wanting toundertakea quantitative analysisof movement. More viable alternatives o taperecording."-.r", are those with built-in memory. This may be in the form ofa hard disc drive (HDD), internal memory (D-RAM) or Flash Memory'

Recording nd storogedeviceA video camera that records the images o tape provides the userwith a numberof options, depending on what type of analysis they are performing. For aqualitative analysis, the recorded movement can be viewed directly from thevideotape in real-time, slow motion or as a still image, using an appropriatevideo flayback system, without the need for any computer hardware orsoftwaie.Alternatively, the user may chooseto capture the video images fromthe tape to a computer hard drive, where they are stored in the form of avideohle (e.g. AVI, MPEG, etc.). This is an attractive option as, with the aidof appropriate software, images can be presented n ways that are not easilyachievabL when playing back directly from tape, for example, the display ofmultiple video clips simultaneously. It also enables a quantitative analysis tobe undertaken, if appropriate digitising, processing and analysis software isinstalled.video images that are recorded to a camera's hard disc drive (HDD),RAM or Flash Memory are usually transferred subsequently to a computerhard drive, where they can be displayed or processed or quantitative analysis.Theprocessof capturing video images o a computer can either be done in real-timeor at some point following the filming session.Which of theseapproachesis akenwill be determined y a numberof factors ncluding he specification

the camera and the filming environment. For video cameras that record toor which have their own hard drive or memory, capture to computerbe done post-recording. I(ith the majority of high-speedcameras this isonly option, as the required data transfer rate exceeds he capability ofryri.-. In most situations with standard 50Hz cameras (and some higher,d cameras),capture of video to computer can be done in real-time. Withlate softwaie, and the requisite connectivity, video sequences rom two

ror. .u-.rus can be captured simultaneously in real-time. When capturing.eomages o a computer, the following practical issuesneed o be considered:

of computerreal-timevideo capture from standard digital videocameras. Firewireas DVin or i-Link). If thishub can be connected ia1394 connection is required (often referred tonot an integral part of the computer, a Firewrre


9/25

16 CARLJ. AYTONt h e U S B o r P C l p o rt . A l t e r n a t i v e l y ' aP C M C I A F i r e w ir e c a r d c a n b e u se d ( f o rlaptops). For some h;;;;;; aigitat camera models' a USB 2'0 ot EthernetpJr, i-r.n-lr.d to do#nto"d "ldtl data' For capture from an,analogue

source'some form of video ."pi"r. u."rd is needed.This *ust be able to capture in afile format "nd resolut'J" inu, is compatible with

the digitising software.All modern."t";;;;;u n."ut "

sufficiently fast processorand adequateR A M t o m a n a g et h e d a t a t r a n s f e r r at e s i n v o l v e d i ns t a n d a r d d i g i t al v i d e ocapture (Firewire ,,lppo"' transfer rates up to 400 Mbitis which is more thanadequate for DV viJeo). An important. consideration is the available harddisc space.The size "ii" ""..r"pressed five second video

file captured at aresolution of 720" szrli".r, i' io.rt 1gMb. Four minuresof uncompressed;;;;;ffi; will therefoie require almost 1 Gb of hard disc space.

Copture sofrworeC a p t u r e s o f t w a r e is u s e d t o c o n v e r t o re n c o d e t h e v i d e ot o t h e r e q u i r e d f il eformat or .codec,.The capture Settlngsused within the software are critical.For quantitur,u. ",.,"i|r;';;;- lo1-1i of. the captured

video file (e.g. AVI'MPEG) must be .#p",iui. with the digirising software. The user should; ; ; ; ; i l " . rp,ur . . in ' t f , t -h igt " " q.ual i tva"vai labie 'mage qual i tv should notbe compromlsed ror the saf,eof fiie size, by using high-compression formats'unless absolutelYnecessary'

VideoploybocksystemAv ideop laybacksys t em is requ i red t o . d i sp l ay t hev i deo i - : q : t f o rqua l i t a t i veor quantiratiu. ".ruryJJ. irr. br,.." should be capable

of displaying 'flickerfree' sti l l images. , ,nt"fa "fst "l lo* video sequenceso be played

in slowmotion and in real-time'

For qualitatiu. """ly'i', an analogue or digital videoplayer-recorder

(VCR) linked ,o " rV- -iriio. i, " ui"Lle option. Thisshould be equipped

with a jog-shuttle dial to control pause and picture advance- unctions' Forana loguev ideo , ro . nu ' S -VHS, a f ou r -headVCn i snecessa ry f o ras t ab l es t i l limage. Some profes,i"""f gt"Jt VCRs will enable individual video fields tobe displayed, ,no, p,luilitig 'ht user with 50 images per second (compared*,rft lT p.. ,..o,td-o" -o'Jdornt'tic video lalers)'Jhit f1:11]lls mportantwhen analysingall but very slow-movemtn"' Tht picturequality on a videomonitor i, irrfl.r.r..jl;;ir.

^;""ri,y .r the source ape, the specification fvideoplaybackdevice,nt typt'of videocableused o link the playbackdeviceto themonito, f n "r.i,tJilg itatt of quality:composite'S-video'Scart

RGB)'.o-fon.",, DVI' HDMI), and the monitor itself'Tradit ionalcnr-"" i , .rsgeneral lyofferexcel lentpictureresolutionbutcannotdirectlydirpl"y ;;i;itul i,r.... LCD Monitors vary in their resolution(e.g.VGA *ont,or, 640 x"480; XGA monitor: 1'024 768; HD monitor:1'366x768).SomemodelsofLCDmonitorscanonlydisplayanalogueSources'-*. ""fy jlgi,"t sources' nd some an display oth'

For qcomputer.card and dby the spesettings,m

Co-ordinTo undertaThis devicthe videobased co-othe still vithat is mawhen selerefers to tsystem is ato whichsystems ofin early sypixel cursoLtd, Coveresolutionlinearly us3 x magni0.05%. TUniversitya digitisingthat, unlesvery 'pixila

DATAC'When confollowed minimise twhen undprocedurerecord of

QuanThe formebeing anamotion. Asubject toappear to


10/25

ro

f

t

MOTION NALYSISSINGVIDEO 17

For quantitative analysis,video playback will be via a laptop or desktopcomputer.^Here,the video data are processed hrough the computer's graphics."ri"rrd displayed on the monitor. The quality of the image will be influencedby the specificationof the graphicscard, the video playback codec' compressionsettings,monitor resolution' and the digitising software'

Co-ordinote igitiserTo undertake a detailedquantitative analysis,a co-ordinate digitiser is required.This device enables wo-dimensional (x, y) co-ordinates of specified points onthe video image, for example, anatomical landmarks' to be recorded. Video-based co-ordiiate digitis.ri ut. essentially software applications that displaythe still video image on a computer screen and overlay this with a cursorthat is manually controlled by the user' The most important considerationwhen selecting a video digitising system is its measurement resolution. Thisrefers to the minimum separation between two points on the screenthat thesystem s able to detect. The digitiser resolution affects the level of precisionto which the co-ordinates can be measured. Current video-based digitisingsystemsoffer considerably higher measurement resolutions than were availablein early systems.This is achieved through a combination of zoom and a sub-pixel cursor. For example, QUINTIC Biomechanics 9.03 (Quintic ConsultancyLtd, Coventry, UK) displays the non-magnified standard video image at aresolution of 720 (horizontal) by 526 (vertical). This resolution can be increasedlinearly using the software's zoom function (up to a maximum of 10x). At a3x magnification, this provides a measufement resolution of approximately0.05%. The TARGET video digitising system developed at LoughboroughUniversity combines a 4x magnification with a sub-pixel cursof to producea digit ising esolution of 12,288 x9,216 (Kerwin,1.995).It should be notedthat, unless he resolution of the captured video is high, the image will becomevery'pixilated' at high magnifications.

DATA OLLECTIONROCEDURESVhen conducting a quantitative video analysis, certain procedures must befollowed carefully, at both the video recording and digitising stages' tominimise he systematicand random errors in the digitised co-ordinates' Even,whenundertaking a qualitative video analysis, many of the video recordingproceduresare still peitinent as they will help to obtain a high quality videorecordof the performance'

Quantitative video analysismay be two-dimensional or three-dimensional.The former approach is much simpler, but it assumesthat the movementbeing analysei- is confi.red to a single, pre-defined plane

- the plane ofmotion. Any measurements taken of movements outside this plane will berbject o perspectiveerror, thus reducing their accuracy. Even activities thatppear o be two-dimensional, such as a walking gait, are likely to involve


11/25

I8 CARL.PAYTONmovements in more than one plane; a two-dimensional analysis would notenable these to be quantified accurately. Three-dimensional analysis enablesthe true spatial movements of the performer to be quantified. This approacheliminates perspective error, but the video filming and analysisproceduresaremore complicated, and the equipment requirements are also greater.

Two-dimensionolideo ecordingThe following guidelines are designed o minimise t he sys tematic and randomerrors present in two-dimensional co-ordinates, resulting from the videorecording stage.This will increase he accuracy of any parameterssubsequentlyobtained from theseco-ordinates.The guidelines are basedon those previouslyreported in Bartlett, 1.997b, and in earlier texts (Miller and Nelson, 1'973;Smith. 1975\.

Equipment set-upMount thecomera on o stoble ripod and avoid ponningThe standard approach in a two-dimensional analysis is for the camera toremain stationary as he performer moves hrough the field of view. This enablesthe movement of the performer to be determined easily relative to an externalframe of reference.

Two-dimensional filming techniques nvolving panning or tracking cam-eras have been usedwhen the performance occurs over a long path (for exampleGervaiset aI.,1989; Chow, 1,993).As thesemethods nvolve he cameramoving(rotating or translating) relative to the externalframe of reference,mathematicalcorrections have to be made for this movement if accurate two-dimensionalco-ordinates are to be obtained.Maxi mi se thecomero-to'subiect distanceThe camera must be positioned as far as is practically possible from theperformer. This will reduce the perspectiveerror that results from movementoutside the plane of performance (seeFigure 2.2).A telephoto zoom lens will enable the camera-to-subjectdistance to beincreased whilst maintaining the desired image size. Note that image qualitywill be reduced if a digital video camera is positioned beyond the limit of itsoptical zoom system.Moximise the image sizeTo increase he accuracy during digitising, the image of the performer must beas large as possible. Image size s inversely proportional to the field of view ofthe camera. The camera should therefore only be zoomed out sufficiently forthe field of view to encompass he performance path, plus a small margin forerror.For events that occur over long performance paths, e.g. triple ;'ump,a single stationary camera would not provide an image size suitable for

(a )Figure2.2to-subiect iswidth oport

quantitativecameras, oFocusthecMost videoverriddenFor a welmanuallyAlign theoAny movenot be subparallel toAs no humof the actican then blens to theof the optorthogonatriangles (optical axhave a det2003). Eveoutside thFigure 2.3


12/25


13/25

20 CARLJ. AYTON

(a) (b )Figure .3 Distortion f ongleswhenmovementccurs utsideheploneof motion. he ruevolueofongles ond B s 90' (imoge ). In mogeb, ongleA oppeorso begreoter hon90" (A')ond ongleBoppeorso be esshon90" (B'),os he rome s no onger n theploneof motion

Record o vertical referenceTo enable a true vertical (and horizontal) frame of reference o be establishedat the digitising stage, a clear vertical reference,such as a plumb line, must berecordedafter the cameraset-up has been completed.Any good video digitisingsystem will correct for a non-vertically aligned camera, using the co-ordinatesof the vertical reference.Record a scoling obiectAn object whose dimensions are accurately known must be recorded in theplane of motion. This is to enable image co-ordinates to be transformed toobject-space real world) co-ordinatesfollowing the digitising stage.Recordingof the scalingobject(s)must be done only after the camera set-up is complete.

The use of both horizontal and vertical scaling objects s essential,becausethe computer may display the image with an aspectratio (ratio of the width tothe height) that distorts it in one dimension. To min imise the error in the scalingprocess, he dimensions of the scaling objectsshould be such that they occupya good proportion of the width and height of the field of view. For a givendigitising error, the scalingerror will be inversely proportional to the length ofthe scaling object. For field widths greater than 2-3 m, scaling is usually doneusing the known distance between two or more referencemarkers or controlpoints, positioned in the plane of motion.

In some circumstances it is not possible to align the camera opticalaxis correctly with the plane of motion, for example when filming in acompetition. Here, digitisation of a grid of control points, placed in the planeof motion, can be used to correct for the cameramisalignment. This methodis called 2D-DLT and has been shown to provide significantly more accuratereconstruction of two-dimensional co-ordinate data than the more commonlyused scaling techniques, particularly when the optical axis of the camera is

tilted moKerwin, 2Se/ecf on aIn activitidistal bodspeedshothe fastesspeed depsuch as a a secondor a swimappropriatspeedof 1An inin image qobtain themust be prexcessivelyEnsurecorrIf filming inlevel. Bart lthe plane oshould provotten prefeinevitably lsun will resa good contrVhen filminpreferred. Vlight source

white balanSelecton opStandard PAeffectively bMost high-sdepend ondependent vdetail) statethat of the hrate shouldhigher).

A suffiand minimumKey events rnswing) are re


14/25

MOTIONANALYSISSING IDEO 21t i l ted more than :r few degrees elative to the pl:rne of motion (Brer,vin ndI (erwin, 2003).Se/ecf on oppropriote shutter speedond opertureIn activit ies such :rs running, jumping, throwing and kicking, it is the rnostdistai body segments, he hands and feet, which move the quickest.A shutterspeedshould be selected hat is sufficient o provide a non-blurred image ofthe fastest novirrgbody segments or sports mplements).Thc choiceof shutterspeeddependson the type of activity being recorded.For slow movernenrs,such as a gr:rnde p1i6 n b:rl letor walking, shutter speedsof 71750-11250 fa secondshould be :rdequate; or moderately ast activit ies, such as rr-rnningor a swimming start, shutter speedsof 11350_11750 f a second are moreappropriate; or fast erctivit ies uch as a golf swing or a tennisserve,a shr-rt terspeedof 1/1000 of a secondor: above rnay be needed .An increarsen shutter speedwill :rlways be accompaniedby a decreasein imarge uality, for given tightingconditions and cirmeraaperturesetting.Toobtain t he best possible magesat the requiredshr-rtter peecl, uff icient ightingmust be proviclcdsuch hat the camera ris aperturedoesnot have o be openedexcessively.Ensurecorrect lighting of the performerIf f i lming indoors, f loocll ightsare often needed o achieve he required ightinglevel. Bartle t, 1997a,sugplestshat one floodlight posit ionedperpendicular othe plane of performance, rnd one to eirch side :rt :rrouncl30' to the plane,should provide ildequate l lumination. Filming outdoors in natural daylight isoften prcferable to f i lming unclerar:t if icial ights, but natr-rral ight levelsareinevitably esspredictable. Jfhcn filming in direcr sunlight, rhe posit ion of thesun will restrictwhere the cirmeracan bc located.The background musr providea good contrast vith the performer and be asplain and unclutterecl s possible.\ff/henfilming indoors with f loodlighting,a dark, non-reflective ackground ispreferred.Vicleocirrner:rs ften havc :.rmanually :rdjust:rblc etting or clifferentlight sor,rrcese.g.daylight, f luorescent irmps, socliumor mercury lemps) .rndwhite balance,which cirn be uscd to enhance he coiour rendit ion.Selecton oppropriote frqme rqteStandarcl'}ALvideo cameras :rve r fired frame retc of 25H2,,although his caneffectivelybc cloubled,proviclccl he c:rmerauses hc interlacedscan method.Most high-spccd irmeras :rvcadjustirble rame rates.The fr:rme ate usedwilldepend on the frequ encycontent of the moverncnt being an:rlysed,ancl thcclepenclcntariablesbeing studied.SamplingTheorum (seeChapter 7 for moredetail) states hart he sarnpling recluerrcyfrarne ate) rnust be at least doublethat of the highest requencyprcsenr n the activity itsclf . In re:rl ity, he fr,rmcr:rtcshould be n'rr-rchigher than this (Challis et al., 1997)suggesr -10 tin-reshieher) .

A sufficicntly high frarne rateirnd minimnm displ:rcement linearkey events n a performancc (e.g.sr , rr rg) rr t ' t 'corded. rr ncrerse n

will ensure hat the instances f maximumar.rd ngr-rlar ) f a joint or l ir nb, and of otherheel-strike n mnning, birl l impact in :r golfthe frame rate will also serve o improve the


15/25

aaaa

Figure 2.4 T(middle ow) t

22 CARLJ. AYTONprecision, and therefore the accuracy, of temporal measurements, or example,the phase durations of a movement. This is particularly important where thephasesare of short duration, for example,the hitting phaseof a tennis serve.Some suggested rame rates for a variety of activities aregiven below:

25-50 Hz - walking, swimming,stair climbing.50-100Hz - running,shotput, high jump.1.00-200F{2 sprinting, avelin hrowing, football kick.200-500 Hz - tennisserve, olf swing,parry in fencing.

It should be noted that these frame rates are only offered as a guide. Fora given activity, the appropriate frame rate should be determined by thefrequency content of the activity and the dependent variablesbeing measured.For example, a quantitative analysis of the interaction between the player'sfoot and the ball during a football kick would require a frame rate above1000 Hz, whereas a rate of 25 Hz would be more than adequate or determiningthe length of the final stride during the approach to the ball. The effect ofusing different frame rates on the recording of a football kick is shown inFigure 2.4.

Participant preparation ond recording triolsThe health and safety of the participant is paramount during any testing.Informed consent should always be obtained from the participant (seeBASESCode of Conduct in Appendix 1) and completion of a health questionnaire isoften required. Sufficient time must be allocated for a warm-up and for theparticipant to become fully familiar with the testing environment and testingconditions.The clothing worn by the participant should allow the limbs and bodylandmarks relevant to the analysisto be seenclearly. The careful placement ofsmall markers on the skin can help the analyst to locate body landmarks duringdigitising, but the positioning of these markers must be considered carefully.Movement of soft tissue means that surface markerscan only ever provide aguide to the structures of the underlying skeleton. Markers are often used tohelp identify the location of a joint's instantaneouscentre of rotation. \Thilst asingle marker can adequately representthe axis of a simple hinge joint, morecomplex joints may require more complex marker systems(seeChapter 3 formore detail on marker systems).The number of trials recordedwill depend on the purpose of the analysisand the skill level of the participants. As the movement patterns of skilledperformers are likely to be significantly more consistentthan those of noviceperformers ('$Tilliams and Ericsson,2005), they may be required to performfewer trials in order to demonstrate a typical performance.During the filmingit is often useful to record a board in the field of view, showing informa-tion such as the date, performer, trial number and condition, and camerasettings.


16/25

@@@reMigure .4 The effecto[ comero rome rote on the recordingof o footboll kick.At 50 Hz (top row] the foot is only seen n contc(middle ow the oot remoins n contoct or four imoges;ot 1000 Hz (bottom ow) the oot s incontoct or sixteen moges notoll


17/25

24 CARLJ. AYTONThree-dimensionolideo ecordingMany of the proceduresdescribed n the previous section for two-dimensionalvideo analysis will also apply when using a three-dimensional approach(selecting an appropriate frame rate, shutter speed and aperture; ensuringcorrect iightitrg of th. performer; maximising the image size and focusingthe camera manually). This section will discuss he main issues hat must beconsideredat the recording stageof a three-dimensional analysis'

Equipmentet-upThe essential requirement is to have two or mofe cameras simultaneouslyrecording the peiformance, each from a different perspective.The choice ofalgorithm or.J to reconstruct the three-dimensional, real world co-ordinatesfrom the two-dimensional image co-ordinates is important as some algorithmsplace severe estrictions on camera locations.

Some three-dimensional reconstruction algorithms rely on very precisepositioning of the cameras relative to one another. For example, the methodproposed Ly Martin and Pongrantz, 1974, requires the optical axes of thei"-..", to be orthogonally aligned and intersecting.Suchmethods can involveconsiderable set-up time and may be impractical to use n some environments,e.g. in sports competitions, as they are too restrictive." The most widely used hree-dimensional reconstructionalgorithm used nsport and exercise biomechanics is the Direct Linear Transformation (DLT)aigorithm. This approach does not require careful camera alignment andth-us allows more flexibility in the choice of camera locations' The DLTmethod determines a linear relationship between the two-dimensional imageco-ordinatesof, for example, a body landmark, and the three-dimensional,realworld co-ordinares of that landmark. A detailed theoretical background to theDLT algorithm can be found elsewhere (e.g. Abdel-Aziz and Katata,"l.971;Miller, Shapiro and Mclaughlin, 1980).To eitablish the relationship between the two-dimensional imageco-ordinatesand the three-dimensional real world co-ordinates,an object spaceor performance volume must be definedusing a setof control points whose real*orld, three-dimensional co-ordinatesare known. This is usually achievedusinga rigid calibration frame of known dimensions incorporating a set of visible-".k.., such as small spheres seeFigures 2.5 and 2.6). Alternatively, a seriesof discretecalibration poles can be used,provided their real world co-ordinateshave been accuratelyestablishedusing, for example, surveying techniques.A minimum of six non-coplanar control points is required for thereconstruction of three-dimensional co-ordinates, but 15-20 control pointsor more is recommended. The control point co-ordinates must be knownrelative to three orthogonal, intersectingaxes' which definea global co-ordinatesystemor inertial refei.nce system.This referencesystem s fixed in spaceandail three-dimensional co-ordinates are derived relative to this. Images of thecontrol points are recorded by each of the cameras being used in the set-up.These are then digitised to produce a set of two-dimensional co-ordinates for

Figure2.5Technologi

Figure .6 C

each contcompute and positobject spafrom camand 2.2.

xr *Yr*

For the mcamera vie


18/25

MOTION NALYSISSING IDEO 25

Figure .5 Colibrotionrome l .60m x 1 91m x 2.23m)with 24 controlTechnologiesnc.)points (PeokPerformonce

Figure.6 Colibrotionrome 1 Om x I .5m x 4.5m)with92 control ointscourtesyf Ross onders)

eachcontrol point from each camera view. These co-ordinatesare used tocompute he 11DLT parameters (C1-C11), which relate to the orientationand oosit ion of each of th. cameras.For control point #1 with real world'object paceco-ordinates Xr, Yr,z1) and with digit isedco-ordinates x1 y1from camera1, the DLT equations or that cameraare given by equations2.1and2.2.

x r * C t X t - f CzYt - l C t Z t ' t C+ Cex1X1 C1ex1Y1 Cnx1Z 1 : Q( 2 . 1 )y1 * C.sXi l C6Y1' t CzZt * Cg+ CsytXr -FCtoytYt I C1:y1Z1 0(2.2)

For he minimum of six control points, twelve equations are produced for eachcamera iew. As thereare more equations han unknowns, the DLT parameters

d


19/25

26 CARLJ. AYTONareob t a i nedbyso l v i ng t heequa t i onsus i nga leas t -Squa res t echn ique (M i l l e r ,*;;;;"J M.Lu,rghii.,, 1980).' fith. the.DLT parametersobtained, the same.q-"u,io", can then U. "r.a to ott"i.t the three-dimensional co-ordinatesof

anymarker in the obiec, ,f".., provided the two-dimensional co-ordinates of themarker are known from at least two of the cameras'when settingup the equipment for three-dimensionalanalysis he biomech-anist should follow ih. tt.pt in the sub-sectionsbelow'

They probeing direMoke proIdeally, thare syncthe cameobtainedsection inDuror strobeto confircorrespoFailure toco-ordina

VideoThe procon theor manuSIMI'Motmarkers option fothe perfoautomatiin envirofilming o

Maidentify awill inevdata (sebe kept tmanuallyr The

cont Onanamuo Grepoithe. Onerrtheima

Mount the comeroson stoble tripods ond avoid ponningThe standard approach in a three-dimensional nalysis s tor the cameras orem a inSt a t i ona ryaSt hepe r f o rm erm oves t h rough t he i r f i e l do f v i ew . T h ree -dimensional filming t..h.riq,r., involving panning camerasJe.g'.Yuet al', 1993;Yanai et at., 1996) ".tJ punni,tg and tilting cameras (e'g' Yeadon' L989)

havebeen used when the p.,fot"tu"tt occurs over a long path' As these methodsinvolve the cameras iorring relative to the global co-ordinate system,a numberof fixed referencemarkers f,uu. ,o be digitised n eachvideo image to correct forthe changing orientation of the camerui. Att alternativemethod of establishingthe orientation of p;; and tilting video cameras was developed by Peakperformance 1...fr.ologi.s" Inc. This ]tuolu., the use of instrumented tripodh;;J, each equipp.d #th two oprical encoders, o sense he angular positionsof the cameras.Positionthe cameros or optimumviewing of body l1nlmolks .Grear care should be taken to ensufe that the body landmarks of interest (e'g'segment endpoints) ,..n"1., in view of at least two cameras or the duration ofthe activity. Inappropriately positioned camerascan result in the analyst havingio g,r.r, li-b poritio* "itt*"1" stagesof the movement' which will inevitably.oipro.rrlr. the accuracyof the co-Jrdinate data' Many video motion analysisp.og'ru--., offer an interpolation function that can predict the co-ordinatesofi^Ulay f""amark that has'becomeobscured.This option should only be usedin situations where the landmark is concealed for no more than four or fiveimagesand is not reaching a turning point (maximum or minimum)

during thatperiod.Ensurecontrol points orevisible to ond recorded by oll comeros.The control points ,rr.J,o compute the DLT parametersmust be clearlyvisibleto .".h ."-",". .srhen using a calibration frame, such as the one shown inFig";. 2.5, careshould b. tluk"tt to avoid the poles at the. rear of the frameL.i"g-"ur."red by those n the foreground (or by the tripod). A good contrastbetween the .o.rtrol point, "nd the

"backgroo.d it also essential' t is advisableto record the control'points at the start and end of the data collection session'This will allow the urrulyrt to recalibrate if one of the cameras s accidentallymoved slightly during the session'Align the performoncewith the axesof theglobol co-ordinate systemThe International iociety of Biomechan-ics (ISB) recommends that, wherethere is an obvious direction of progression, for examp-le n. gait, the X-axiso f t h e g l o b a l c o . or d i n a t e , y , , . . b e n o mi n a l l y a l ig n e d w i t h t h i s d ir e c t i o n .


20/25

MOTION NALYSISSING IDEO 27They propose the use of a right-handed co-ordinate system,with the Y-axisbeing directed vertically and the Z-axis laterally.Moke provision for shuftersynchronisation ond event synchronisotionIdeally, the two or more camerasshould be gen-locked to ensure heir shuttersare synchronised. Where this is not possible, the time lag between each ofthe camera shutters must be determined so the two-dimensional co-ordinatesobtained from each camera view can be synchronised (seeVideo Camerassection n this chapter).During filming, it is useful to activate an event marker e.g. an LEDof strobe light that is visible to all cameras. Such a device can be usedto confirm that the first video image digitised from a given camera viewcorresponds emporally to the first image digitised from all other camera views.Failuri to fulfil this requirement will result in erroneous three-dimensionalco-ordinates.

Video igitisingThe process of obtaining two-dimensional co-ordinates of specified andmarkson th. performer, from a video record, may be achieved automaticallyor manually. Most video motion analysis systems (e.g. APAS, Qualisys,SIMI'Motion) now include software that can automatically track passivemarkers affixed to the performer. While this facility is clearly an attractiveoption for the user, it is not always possible or practical to place markers onthe performef, e.g. during a sports competition. Even where this is possible,automatic tracking of passive markers can still be problematic, particularlyin environments where the contrast level of the marker is variable, e.g. whenfilming outdoors or underwater.Manual digitising of a video record requires the biomechanist to visuallyidentifyand mark the anatomical sitesof interest, ftame-by-ftame.This processwill inevitably introduce some systematicand random errors to the co-ordinatedata (seeChapter 7 for more detail). Sfith attention to detail, theseerrors canbekept to an acceptable evel.The following points should be consideredwhenmanually digitising a video sequence:o The same operator should digitise all trials in the study to ensure

consistency(reliability) between trials.. Only ever use skin-mounted markers as a guide. Consider carefully theanatomical landmark being sought. A sound knowledge of the underlyingmusculo-skeletal system s essentialhere.o Great care should be taken when digitising the scaling object or controlpoints. Any measurement error here will introduce a systematicerror inthe co-ordinate data.and in all variables derived from these.

o On completion of a 3D calibration, check that the 3D reconstructionerrors fall within acceptable imits. These errors will depend mainly onthe volume of the object-spacebeing calibrated, the quality of the videoimage and the resolution of the digitiser. As a guide, Sanderset al',2006,


21/25

I8 CARLJ. AYTONreportedmean RMS reconstructionerrors of 3.9mm, 3.8 mm and 4.8 mmfor the x, y and z co-ordinates, respectively, of 30 points distributedthroughout the large calibration frame shown in Figure 2.6.o I representative sequenceshould be digitised several imes by the inves-tigator to establishthe intra-operator reliability. Inter-operator reliability(objectivity) should also be determined by having one or more otherexperienced individuals digitise the same sequence(see Chapter 7 andAtkinson and Nevill. L998. for more information on assessingmeasure-ment reliabil itv).

PROCESSING,NALYSINGNDPRESENTINGVIDEO-DERIVEDATAThe video digitising process creates two-dimensional image co-ordinates thatare contaminated with high frequency errors (noise). Essentially, what isrequired next is to: 1) smooth and transform the co-ordinates,so that they are na form suitable for computing kinematic variables,2) calculateand display thekinematic variables in a format that allows the user to extract the informationrequired to complete the analysis.

Smoothing nd tronsformingo-ordinotesThere are various smoothing methods that can be used to remove the highfrequencynoise ntroduced by the digitising process; hese all into three generalcategories:digital filters, spline fitting and fourier series truncation (Bartlett,1997a). Failure to smooth co-ordinates sufficiently will lead to high levelsof noise in any derived kinematic variables, particularly acceleration. Over-smoothing of the co-ordinates will result in some of the original signal beinglost. Selecting the correct smoothing factor, for a given set of co-ordinates,is therefore critically important. Chapter 7 provides a detailed discussion ofsmoothing methods and presents some practical guidelines for their use. Thetransformation of image co-ordinates to real world co-ordinates is necessarybefore any analysis can be undertaken. Procedures for achieving this werediscussedearlier in this chapter.

Colculotig kinemoticorioblesThe sport and exercisebiomechanist is often interestedboth in the movementpatterns of individual body segments, or example in throwing and kicking'and in the overall motion of the performer's centre of mass, for example in asprint start. Computation of the mass centre location requires a linked-segmentmodel to be defined, and the mass, and mass centre locations, of individualbody segmenrs o be determined. Three general methods are used to obtain

body segfrom cadimaging din Robertdata thatanalysed

The dimensioco-ordinatlinear disPythagoradimensionrelative (ethe anglesimple to system hathree-dimemost comin biomecA detailedLinesecond timThese dermethod) ofunctions)and accefound.

AnolysiIn any biodeterminedvariablesas this wihigh-speedeterminismovementliterature.

Therevideo analpresentatioof the infocommon mpeak jointthe focus oplots and ain sport an


22/25

MOTION NALYSISSINGVIDEO29body segment parameters: regression equations basedon measurements akenfrom cadavers, geometric modelling of the body segments,and the use ofimaging devices(Bartlett, 1'997a).More detail on thesemethods can be foundin Robertson,2004. The biomechanist should seek to use segmental inertiadata that closely match the physical characteristicsof the participants beinganalysed.

The linear displacement of a body landmark (or mass centre) in onedimension (e.g. r direction) is defined as the change in the relevant scaledco-ordinate of that landmark (Ar) during a specified time period. Resultantlinear displacements in two or three dimensions are easily calculated usingPythagoras' heorem. Two-dimensional (planar) anglesare obtained from two-dimensional co-ordinates using simple trigonometry. These angles may berelative (e.g. joint angles formed by two adjacent segments)or absolute (e.g.the angle of a segment relative to the vertical). Planar angles are relativelysimple to interpret, once the angular conventions adopted by the analysissystem have been established.The calculation of relative (joint) angles fromthree-dimensionalco-ordinates is more complex, as is their interpretation. Themost common methods used for calculating three-dimensional joint anglesin biomechanics are the Euler and Joint Co-ordinate System (JCS) methods.A detaileddiscussionof thesemethods s provided by Andrews,1995.Linear and angular velocitiesand accelerationsare defined as the first andsecond ime derivatives of the displacement (linear or angular), respectively.These derivatives can be computed either numerically (e.g. finite differencemethod) or analytically (if the data have been smoothed with mathematicalfunctions). As with displacement, the orthogonal components of velocityand acceleration can analysed separately, or their resultants can befound.

Anolysingnd presentingideo-derivedotoIn any biomechanical analysis, the selection of dependent variables will bedeterminedby the aim of the study. It is important that the biomechanicalvariables of interest are identified before undertaking the data collection,as this will influence the methodology used (e.g. 2D vs' 3D; normal vs.high-speedvideo). ' (hen analysing a sport or exercise activity, the use ofdeterministicmodels (Hay and Reid, 1982) can help to identify the importantmovementparametefs, as of coursecan reference o the appropriate researchliterature.

There are a number of ways of presenting the kinematic data from avideoanalysis and it is for the individual to decide on the most appropriatetion format. This will be dictated mainly by the intended destination

the information (e.g. research journal, athlete feedback report). The mostmethods of presenting kinematic data are as discrete measures (e.g.joint angles) and as time seriesplots (e.g. hip velocity vs. time). Wherefocus of the analysis s on movement co-ordination, the use of angle-angleand angle-angular velocity (phase)plots is becoming increasingly popular

soort and exercise biomechanics.


23/25

30 CARL. PAYTONREPORTINGVIDEO OTION NALYSISTUDYThe biomechanist should consider including some or all of the followinginformation when reporting a video-basedstudy.

Porticiponts. Participant details (age,height, body mass, trained statusetc.);o Method of obtaining informed consent (verbal or written);o Nature of the warm-up and familiarisation;' Type of clothing worn, type and position of skin/other markers and themethod of locating body landmarks.

Video recordingo camera and lens type (manufacturer and model) and the recordingmedium, format and resolution (e.g.HD 7Z0i on to miniDV tape);. Camera settings (frame rate, shutter speed, ris (f-stop) setting);o Position of camera(s) relative to the movement being recorded and thefield width obtained from each camera (a diagram is useful here);o Method usedto synchronise he cameraswith each other (and with otherdata acquisition systems f used);r Details of lighting (e.g.position of floodlights);o Dimensions of 2D scaling object(s)or 3D performance volume (includingnumber and location of control points).

Videodigit ising' Digitising hardware and software (manufacturer and model/version):. Resolution of the digitising system;. Digitising rate (this may be less han the camera's frame rate):. Model used (e.g. 15 point segmental).

ACKNI wouldphotograMark Joh

REFEAbdel-Azitor co-oAmericaFallsChAndrews,Jthree-diThree-dAtkinson,Gerror (re2r7-238Bartlett, RE & FNBartlett,R.Leeds:BrBrewin,M.techniquChallis,J.,R.M. BaBritishAChow, I ,U/characte9 : 149 -1

Gervais,P.,and KuipCanadianGreaves,J.OP. Allard,MouemeHay, J.G. aMotion, EKerwin,D.G(ed.) ProcSportandMartin, T.PperspectMiller, D.I.PhiladelpMiller, N.R.tial kinemdata',Jou

Processing,nolysing nd reportingaOaaa

Algorithm used to obtain the 3D co-ordinates;Method used to smooth/filter the coordinates;Level of smoothing;Method used to obtain the derivative data (e.g.numerical, analytical);Sourceof segment nertia data used to calculate e.g.the whole body masscentreor moment of inertia:Definitions of the dependent variables being quantified, including their SIunits;Estimation of the measurementerror in the calculatedparameters;Level of inter- and intra-observer reliability of the calculatedparameters.

aa


24/25

MOTIONANALYSISSINGVIDEO3I

ACKNOWLEDGEMENTI would like to thank Ed Parker for his help in preparing some of thephotographsn this chapterandhis technical dvice. would also ike to thankMark Johnson or providing high speed ideo ootage or Figure2.4.

REFERENCESAbdel-Aziz, .I. andKarara,H.M. (L971'l'Direct inear ransformationrom compara-tor co-ordinatesnto object space o-ordinatesn close angephotogrammetry',nAmerican Societyof PhotogrammetrySymposiumon CloseRangePbotogrammetry,Falls Church, VA: American Societyof Photogrammetry.Andrews, .G.(1995) Euler'sandLagrange's quationsor linked igid-bodymodelsofthree-dimensionaluman motion', in P. Allard, I.A.F' Stokes nd J-P.Blanchi eds)Tbree-dimensionalnalysisof buman mouement,Champaign, L: Human Kinetics.Atkinson,G. and Nevill, A.M. (1998) Statisticalmethods or assessing easurement

error (reliability) n variables elevant o sportsmedicine',SportsMedicine,26(4);21.7-238.Bartlett, R.M. (t997al Introduction to Sports Biomechanics,1st edn, London:E & FN Spon.Bartlett,R.M. (ed.) 1997b)Biomechanical nalysis f Mouementn SportandExercise,Leeds:British Association f Sportand Exercise ciences.Brewin,M.A. and Kerwin, D.G. (2003) Accuracyof scalingand DLT reconstructiontechniquesor planarmotion analyses',ournalof Applied Biomechanics, 9:79-88.Challis, ., Bartlett, R.M. and Yeadon,M. (t997) 'Image-basedmotion analysis',nR.M. Bartlett ed.)Biomecbanical nalysis f Mouementn SportandExercise,Leeds:BritishAssociation f Sportand Exercise ciences.Chow,J.W. (1,993) A panning videographicechnique o obtain selected inematiccharacteristicsf the strides n sprint hurdling', Journal of Applied Biomechanics,9:1.49-1.59.Gervais, .,Bedingfield,.V.,'Vfronko,C., Kollias, ., Marchiori, G.,Kuntz,J.,'Way,N.and Kuiper, D. (1989) 'Kinematic measurementrom panned cinematography',Canadian ournalof SportsSciences,14: 07-1"1'1.Greaves,.O.B.(1995) Instrumentationin video-basedhree-dimensionalystems',nP. Allard, I.A.F. Stokes ndJ-P.Blanchi edslThree-dimensionalnalysisof HumanMouement,Champaign,L: Human Kinetics.Hay, J.G. and Reid, J.G. (1982) Tbe Anatomicaland MecbanicalBasesof HumanMotion, Englewood Cliffs, NJ: PrenticeHall.Kerwin,D.G. (1995) Apex/Targethigh resolution ideodigitising ystem'n J. \Tatkins(ed.l Proceedingsof the SportsBiomechanicsSection of the British Association ofSportandExercise ciences,eeds: ritishAssociation f Sportand Exercise ciences.Martin,T.P. and Pongrantz,M.B. (1974\ Mathematicalcorrection or photographicperspective rror', Resed.rh Quarterly, 4 5':378-323.Miller, D.I. and Nelson, R.C. (1973)Biomechanics f Sport: a Research pproach.

Philadelphia, A: Lea& Febiger.Miller,N.R., Shapiro,R. andMclaughlin, T.M. (1980) A techniqueor obtainingspa-tial kinematicparameters f segments f biomechanical ystemsrom cinematographicdata', ournalof B ome hanics,'J-3535 547


25/25

32 CARL.PAYTONRobertson, D.G.E. (2004) 'Body segment parameters', n D.G'E. Robertson,G.E.Caldwell,J. Hamill, G. Kamenand S.N. Whittlesey eds)ResearchMethodsin Biomechanlcs, hampaign,L: HumanKinetics.Sanders, ., Psycharakis, ., McCabe' C.' Naemi' R., ConnaboY'C., Li, S., Scott,G.and Spence, . (2006) Analysisof swimming echnique: tate f the art:Applicationsand implications',Portugueseournalof Sport Sciences,6,upl.2:20-24'Smith,G. (1975) Photographicanalysisof movement', n D. 7. Grieve,D.I. Miller,D. Mitchelson, .P.Paul andA.J. Smith(eds)Techniquesor theAnalysisof HumanMouement, ondon:LepusBooks.Williams, A.M. and Ericsson,K.A. (2005) 'Someconsiderations hen applying theexpertperformance pproachn sport',Human MouementScience,4: 283-307.'Winter,D.A. (1990)Biomechanicsnd Motor Controlof HurnanMouement,2ndedn,New York: Wiley.Yanai, T., Hay, J.G. and Gerot, J.T. (t9961 'Three-dimensionalvideography ofswimmingwith panningperiscopes',ournalof Biomechanics,29:73-678.Yeadon.M.R. (1989) A method or obtaining hree-dimensionalata on ski jumpingusingpanand ilt camer s' Internationalournalof Sports iomechanics, 238-247Yeadon,M.R. andKing, M.A. (t999)'A method or synchronisingigitised ideodata',

Journal of Biomechanics,2: 983-986.yu, B.,Koh,T.J. andHay,J.G.(1993) A panningDLT procedureor three-dimensionalvideography', ournalof Biomechanics,6: 741'-7 1.

CHAP,iiiE1:,' ,

MOONClare

INTRBiomechof threemotion acise.As opefatornot onlyquality dour knowbiomechof furthepreventiby technbetweenrationaletechnolo

Onto reseaof theseof injurysustaininthat remand beinrunningpopularsport. Rits repetare subtbiomech

2 video analysis

Documents