A tracking algorithm for rigid point-based marker models

Abstract

Tracking of objects and persons by using a stereo camera system setup and markers is applied since several years in Virtual and Augmented Reality or other applications. In especial infrared tracking systems are getting increasingly popular, because of their precision and robustness. Often, markers for such systems are spherical and therefore not distinguishable, making the consideration of geometric constellations of several markers necessary to identify objects and to determine their transformations. This paper presents an algorithm for tracking rigid constellations of markers, which can be adapted to the needs of corresponding applications.