Study of color fiducial based registration algorithm for augmented reality

Augmented Reality system augments the real world in the user's view with the information generated by computer. Fiducial marker based registration is the most commonly used method and gray image is often used in the current system which lost all color information. A color fiducial marker based tracking algorithm and relative Gauss-Newton iterative registration system was proposed. The proposed tracking and registration system could avoid the lost of color information in the grabbed images to achieve highly accurate and effective registration. Experiment shows that the accuracy of the proposed system can satisfy the requirement of the existing AR applications.