Effects of the calibration procedure on the metrological performances of stereophotogrammetric systems for human movement analysis

This paper proposes a methodology for evaluating the effect of different stereophotogrammetric system calibration procedures on the calculation of marker-based kinematics information. The methodology, based on calibrating the system using data recorded from capture volumes of different sizes and in trials of different durations, was applied to two different systems. The calibration data were used to reconstruct the static and dynamic position and orientation in space of a rigid wand carrying markers in known positions. The inaccuracies in the reconstruction of distances and angles from the wand markers were independent on the calibration data, with average errors lower than 1.7 mm and 0.7°, respectively. Similar results were obtained from human gait data, with the highest variations observed in the transverse plane kinematics and in the foot segment, suggesting that successful calibration procedures of different durations and performed in different volumes did not affect the metrological performance of the investigated systems.