Validation of a multi-camera videogrammetry approach for quantifying helmet impact velocity in football

This study presents the validation of a multi-camera videogrammetry approach for measuring football helmet impact velocities, wherein the effect of camera angle, camera distance and impact speed were assessed. Helmet-to-ground impacts were simulated within two zones on a field using free fall drop tests of a helmeted head and neck assembly at heights of 1.04 m (slow: 4.52 m/s) and 1.83 m (fast: 5.99 m/s). Helmet motion was manually tracked using 3D motion analysis software for different camera combinations (e.g., orthogonal, coincident, overhead, and parallel camera pairs) in each zone. Maximum errors of the 3D calibrations were < 3 cm; mean maximum uncertainty ( ±) error for 3D tracking was ≤ 2.4 cm across all test conditions, except parallel angles (4.3–10.0 cm). Helmet impact velocities from orthogonal, coincident and overhead angles compared favorably to the reference velocities (slow: 4.50 ± 0.05 m/s; fast: 6.09 ± 0.06 m/s) with relatively low errors overall (relative error (RE), absolute error (AE): ≤ 3.39%; root-mean-square error (RMSE): ≤ 0.22 m/s). Parallel angles had the highest errors (RE, AE: 3.79–10.94%; RMSE: 0.18–0.55 m/s) due to inaccuracies with horizontal measurements in-line with the camera views. In general, increases in camera distance and impact speed did not appreciably influence video tracking accuracy.