Some effects of camera placement on the accuracy of the Kinemetrix three-dimensional motion analysis system

Objective. To assess the accuracy of the Kinemetrix motion analysis system to measure horizontal movement by a single reflective marker away from two cameras with differing camera placements. Design. Trial of the effects of nine different camera arrangements on precision of non-human movement. Background. In many cases the ability to arrange cameras to allow a separation of 60 ° is not possible. Little is known about the precision of motion analysis systems for small camera angle separations. Methods. The accuracy of the Kinemetrix system was assessed with camera horizontal separations of 15 °, 30 ° and 45 °, and vertical separations of 0 °, 15 ° and 30 ° rendering nine different camera placements. The distance between the cameras and the object was always maintained at 4 m. During each test the marker was moved a known horizontal distance along a line bisecting the horizontal angular separation of the two cameras. The mean absolute errors of the Kinemetrix measurement were calculated. Results. At the smallest camera separation tested (15 ° horizontal, 0 ° vertical), the Kinemetrix was unable to calculate the three-dimensional co-ordinate of the marker. For all other camera positions tested, the errors in measurements were small (mean absolute errors <2 mm). Conclusions. Maintaining camera horizontal and vertical separations above a sum of 30 ° is sufficient for clinical testing.