Abstract
The purpose of this study was to investigate the effect of different camera orientations on variable results from 3-dimensional video motion analysis. The analysis included digitised data from 4 cameras, which were located symmetrically around the activity (sprint hurdle clearances). After the digitisation, data files for all 4-camera combinations (4C) and for the four different camera pairs were created. Each camera pair included exactly the same calibration and digitisation information from the respective camera views as the 4C-files. The variable results from 4C were compared with different camera pairs. Factorial ANOVA with LSD post hoc test revealed for three different athletes 16, 15 and 16 variables out of 26 in which camera pairs differed from 4C. The camera pair on the athletes' trail leg side yielded the least statistically significant differences to 4C. One reason for this could be that the joint landmarks on the trail leg, which are crucial for certain variables, are visible all the time for these cameras. The camera pair at the front side of the clearance revealed the most statistically significant differences. The main reason for this could be that these camera orientations may have problems with the deepness of the picture in relation to running direction. This study has indicated that the camera orientations could have a significant influence on variable results. Thus, it is important to consider camera locations carefully, when studying human movement.
| Original language | English |
|---|---|
| Pages (from-to) | 253-267 |
| Number of pages | 15 |
| Journal | Journal of Human Movement Studies |
| Volume | 47 |
| Issue number | 3 |
| Publication status | Published - 2004 |
| Externally published | Yes |
Keywords
- Biomechanics
- Factorial ANOVA
- Multiple cameras
- Sprint hurdles
- Videotaping