Research And Application Of Mobile Augmented Reality's Tracking And Registration Technology

Augmented reality is a comprehensive technology that enhances or expands the real scene by means of the computer-generated information.It is a derivative of virtual reality and a hotspot in recent years.With the development of mobile computing technology,the improvement of mobile hardware performance and the widespread use of smart phones,it will be possible to apply augmented reality into smart phones.Mobile augmented reality is to apply augmented reality into the mobile device effectively,and mobile devices become a media to achieve the interaction between virtual and reality.Tracking and registration technology is the basic technology in augmented reality,as well as the key to the augmented reality system performance.Traditional tracking and registration algorithm based on computer vision is computationally complex and time-consuming.Compared with the computer,the current smart phone's memory is inadquete,and processing capacity is low,thus it is beyond its capability to run the complex algorithm.The accuracy and real-time performance of mobile augmented reality can not be compared with the PC.In this dissertation,aiming at the performance and function requirements of mobile augmented reality,making compared experiments of various feature matching algorithms,compared advantages and disadvantages of these algorithms,the ORB algorithm is chosen as the feature matching algorithm of mobile augmented reality system.In-depth analysis of ORB algorithm,aiming at the problem that the feature points detected by the ORB algorithm are sensitive to scale changes,submit to improve the ORB by combining scale space.In the feature point matching and pose tracking stage,the traditional ORB algorithm uses the exhaustive method to carry on the feature matching,thus the computational time is huge,does not apply to the mobile phone.Therefore using the bad-of-word model to change the feature search strategy,thus speed up the feature point matching efficiency,calculate the camera posture,to achieve real-time tracking of mobile phone pose.A mobile augmented reality system is designed,and a "recognition + tracking" mobile augmented reality framework is proposed by means of applying the improved algorithm into the system.So as to improve the robustness and real-time performance of the mobile terminal algorithm,and this mobile augmented reality is suitable for large-scale indoor and outdoor scenes.