Quantifying Rotations of Spheric Objects

This work presents an approach to quantify rotations of spheric objects with a single camera. We apply this method to objectively classify racket sports equipment. Therefore, we observe the ball impact on a racket and compare differences in rotations quantified prior to and after the impact. Our approach combines motion segmentation with tracking both the ball center and corners of the surface texture. A 2D warping of these corner positions reveals the rotation. We neither require a motion model nor any user intervention. Experimental results verify our approach and prove the feasibility of our application scenario based on visual rotation quantification.