RECENT PRACTICAL APPLICATIONS OF CLOSE-RANGE PHOTOGRAMMETRY 
FOR COMPLEX MOTION STUDY 
Albert K. Chong a , Peter Milbum b , Richard Newsham West b , Marieke ter Voert b , Hayden Croft' 
School of Surveying, University of Otago, Dunedin, New Zealand - albert.chong@surveying.otago.ac.nz 
School of Physiotherapy, University of Otago, Dunedin, New Zealand 
School of Physical Education, University of Otago, Dunedin, New Zealand 
Commission VI, WG VI/4 
KEY WORDS: Photogrammetry, Spine study, Rugby lineout throw, 3D virtual training system, Stereovideo capture, Body 
segment movement. 
ABSTRACT 
This paper discusses two practical applications of close-range photogrammetry completed at the University of Otago recently. The 
applications involve: 1) human spine study; and 2) rugby player’s limb coordination. The discussion covers: 1) imaging system 
design; 2) system calibration; 3) image processing and 3D data capture; and 4) usage of the captured 3D data. Also, the spatial 
accuracy requirement of these applications and the potential of the close-range photogrammetric technique for these areas of research 
are highlighted in the paper. Results of the projects show that digital photogrammetry is efficient and accurate for the applications 
discussed. 
1. INTRODUCTION 
This paper introduces two recent applications of close-range 
photogrammetry for the study of the human body in relation to 
spinal movement and rugby sport players’ limb coordination. 
The main measurements are distances, vectors and angles. The 
measurements were needed to evaluate the parameters required 
for their respective studies. Identifiable points or anatomical 
landmarks within the region of the spine or the limbs are 
required to obtain accurate measurements. Still-frame/video 
imaging and laser technology are the best known methods for 
these types of study. Recent works on human body 
measurement using video can be found in Figueroa et al. 
(2003). By and large, hardware and software are readily 
available off-the-shelf for human motion analysis (Richards 
1999). The author reported RMSs of seven major systems were 
less than 2.0 mm in the RMS. 
LANC Shepherd (Graphic Media Research, Canon Falls, Mn 
US) remote camera control device were used to synchronise the 
shutters of the cameras. The cameras were pre-calibrated 
before images of the patients’ spines were captured for the 
project. To calibrate the cameras, images were taken with a set 
of temporary retro-targets (Figure 1) fastened onto the control 
frame and an exterior orientation device as shown in Figure 1. 
2. METHODOLOGY 
2.1 Spinal study 
In the human spinal study, the spinal movement and spinal 
deformation of athletes were needed to evaluate the occurrence 
of spinal injury in a gruelling training environment such as 
military exercises in rough terrain. Generally, in the laboratory, 
the high performance imaging system such as the Motion 
Analysis (Motion Analysis Corporation, Santa Rosa, CA) 
system provides accurate 3D data for motion study. However, 
the portability and workability of such a system is limited in a 
harsh outdoor environment. A portable image-capture device 
was developed to enable in-situ data capture for the study. The 
device consists of a sturdy aluminium frame which can support 
four cameras, a rigid control frame, light-fitting and two 
mounting-brackets for a surveying tripod. The four cameras 
could take either convergent photography or stereophotography 
of the whole spinal region simultaneously. The device could be 
mounted on two surveying tripods for photography in the field. 
Four Sony Cyber-shot digital cameras and a four-camera 
Figure 1. A convergent image used in the system calibration. 
Note the temporary retro targets (inside the control frame) used 
for lens distortion calibration. 
In the system calibration, the spine anthropometric landmarks 
(cervical, thoracic, lumbar, and pelvic) positions were marked 
by an experienced physiotherapist (Figure 2). Multi-convergent 
and stereophotography were obtained of the participants’ backs 
at predetermined time intervals throughout the day. 
Additionally, the framing mode of capturing images was used 
where the movement must be tracked continuously for a short 
period. Australis, photogrammetric bundle software 
(Photometrix Pty Ltd, Melbourne Australia) was used to 
calibrate the camera on-the-job as the handling of the camera 
could have altered lab calibration. VirtualDub 
(WWW.VirtualDub.org) image processing software was used 
for the image process and Australis was used to carry out all 
the bundle adjustment of the images. The results of the camera 
921 
im 
I 
I 
Mils 
Mm 
iS"