3D trajectory reconstruction of the soccer ball for single static camera systems

The interest on acquiring player and ball data during soccer games is increasing in several domains such as media or training. Consequently, tracking systems are becoming widely used for live data gathering. However, due to costs, stadium infrastructure, media rights etc. there is a trend for stand-alone mobile low-cost soccer tracking systems. The drawback of such systems is that generally only low-resolution images of the players are available. Thus, the problem of detection and tracking the soccer ball is also strongly exacerbated. This paper presents a novel approach for 3D reconstruction of the ball trajectory from monocular lowresolution soccer image sequences. The ball detection is done on accumulated motion-segmented binary images by extracting shape-validated ball candidates from these so-called motion history images. Then, robust ball tracklets are extracted from the motion history images and used to reconstruct the ball 3D trajectory based on physical characteristics and calibration information. The approach is tested on a Bundesliga data set: the tracklet extraction and the ball trajectory reconstruction are evaluated.