Fast reconstruction of centre of mass and foot kinematics during a single-legged horizontal jump: A point-cloud processing approach

Abstract

Horizontal jumps are discrete, fast, over-ground movements requiring coordination of the centre of mass (CoM) and base of support and are routinely assessed in sports settings. There is currently no biomechanics-based system to aid in their quick and objective large-scale assessment. We describe a practical system combining a single low-cost depth-sensing camera and point-cloud processing (PCP) to capture whole-body CoM and foot kinematics. Fourteen participants performed 10 single-leg horizontal jumps for distance. Foot displacement, CoM displacement, CoM peak velocity and CoM peak acceleration in the anterior-posterior direction of movement were compared with a reference 15-segment criterion model, captured concurrently using a nine-camera motion capture system (Vicon Motion Systems, UK). Between-system Pearson’s correlations were very-large to near-perfect (n = 140; foot displacement = 0.99, CoM displacement = 0.98, CoM peak velocity = 0.97, CoM peak acceleration = 0.79), with mean biases being trivial–small (-0.07 cm [0.12%], 3.8 cm [3.5%], 0.03 m·s−1 [1.6%], 0.42 m·s−2 [7%], respectively) and typical errors being trivial–small for displacement (foot: 0.92 cm [0.8%]; CoM: 3.8 cm [3.4%]) and CoM peak velocity (0.07 m·s−1 [4.3%]), and large for CoM peak acceleration (0.72 m·s−2 [15%]). Limits of agreement were −1.9 to 2.0 cm for foot displacement, −11.3 to 3.6 cm for CoM displacement, −0.17 to 0.12 m·s−1 for CoM peak velocity and −2.28 to 1.43 m·s−2 for CoM peak acceleration. The practical system captured CoM and foot kinematics during horizontal jumps with acceptable precision. Further work to improve estimates of CoM accelerations and different populations are warranted.