| Augmented reality (AR) is the technology of merging synthetic sensory information into users'perception of a real environment. The principles and implementation of the registration technology in AR are covered in this dissertation.The registration technology, as the critical technique of AR, is the main problem to be faced with AR. The task of registration is to determine the mapping relationship of the virtual objects between the different coordinate spaces, and the virtual objects are augmented on the real scene in the proper positions in soft real-time. In recent years, with the rapid development of computer graphics and computer vision, the registration technology based on computer vision shows great advantages, such as simpler equipments, lower costs, and results in better real time and more accurate registration.In order to fulfill the AR system based on computer vision, a registration method using the projective reconstruction is proposed. Two steps of the registration method are included: embedding and tracking. In the embedding step, four points are specified to build the world coordinate system so that the virtual objects are superimposed on the real scene images in the proper positions. During the course of tracking, the projective reconstruction is used to track the four points and to calculate the transformation of model views. On the basis of the camera calibration, implementation principles are introduced, and details of the embedding process are given and analyzed. To implement the features of the better real time and more accurate registration, the manual vision marks of black-white pattern are designed. Next, the captured video images are transformed into binary images, feature points are detected, and feature points are tracked using tracking technology then. In addition, a series of experiments and tests related are designed to study the registration method. Tests show that the registration is well done and has the advantages of better accuracy and high speed. |
PDF Full Text Request