Linear CCD Based Optical Tracking Using Stereo Correspondence Algorithm

Optical Tracking is an important part in the computer aided surgery system. Previously, the tracking of interaction devices with multiple retro-reflective markers can only be performed by two-dimensional (2D) image sensors, however with an inferior positioning accuracy compared to the tracking system based on one-dimensional (1D) image sensor. In this paper, we study the system based on 1D sensor from the point of view of computer vision, and present our novel algorithm that solves the correspondence problem in the simultaneous positioning of multiple markers with this system. Experiments are designed to implement this algorithm to track an interaction device with four infra-red markers lighting at the same time. Results show that this algorithm can effectively identify the correspondence of multiple markers, thus break through the limitation that 1D image sensors can only position one marker at a time.