Three-Dimensional Foot Contact Position on a Smart Fitness Trampoline with a Upward Looking Wide View Imaging System

Abstract

This study estimates the position and depth of the user s foot contacts of playing on a trampoline by acquiring the image of the footprint shadow and processing it to enable quantitative analysis of jumping (as a fitness exercise) or to link it to computer games. First, a hardware prototype of a jumping fitness monitoring system is proposed, consisting of a trampoline, an upward-looking wide-angle fish-eye camera module, and a signal processing embedded board. Second, image-based three-dimensional foot contact position estimation (I3D-FeE) algorithms are proposed, comprising the foot contact position estimation algorithm in the transverse plane and the foot contact depth estimation algorithm in the vertical direction. The foot contact positions were estimated using the binary image of the foot shadow obtained by extracting the binary masks of the trampoline and foot. The foot contact depth was estimated using the similarity ratio of the footprint shadow sizes. To verify the usefulness of the proposed system and the accuracy of the proposed contact position estimation algorithms, a series of experiments were conducted using a robotic manipulator. The experimental results show that the X and Y coordinates of the foot contact position and the Z coordinate of the foot contact depth were successfully estimated within an average error of 3.40 (mm), 3.26 (mm), and 4.49 (mm), respectively.