Research In Mobile Augmented Reality System For Outside Architectural Environments

In recent years,Augmented reality(AR)has become a new way of geographic expression of geographic information system.It is a technology that users can observe the real world by means of equipment,adding virtual information generated by computer such as multimedia objects and 3D models.With the rapid development of mobile devices,the domestic and overseas experts and scholars have done a lot of research and practice on AR,but most of them are focus on the PC side.Compared to the PC terminal,the mobile terminal has greatly expanded the application scene of the AR system because of its portability,and users can have more interaction with the real world.However,due to the limited computing capacity of mobile terminals,the mature algorithm,which has been widely used in computers,can not be directly transplanted to the mobile terminal,resulting in the existing algorithm can not meet the real-time requirements of outdoor registration method.Aiming at the advantages and disadvantages of augmented reality in the mobile terminal application,the focus of this paper is to study the real-time and robustness of the registration method of mobile augmented reality systems in outdoor scenes,and analyzes the key technologies,including image matching technology and tracking registration technology.On the research of image matching technology,this paper proposes an improved SUFREAK algorithm which combined with SURF algorithm and FREAK algorithm.Using mobile terminal as the experimental platform,from the aspects of matching performance and rotation,illumination,scale,perspective transformation makes comparative analysis on algorithm efficiency and robustness,and to explore the applicability of the algorithm in the mobile terminal from the comprehensive perspective.The results show that the average time-consuming of the improved algorithm SUFREAK is about 49.79%of the SURF algorithm,which overcomes the shortcoming of the real-time nature of the SURF algorithm,and its robustness is better than the SURF algorithm to a certain extent under the perspective and illumination conversion.Meets the robustness requirements on mobile terminals.In order to fully exploit the temporality between video frames to ensure the fluency of the enhanced system and the efficiency of the target tracking,a registration method based on the SUFREAK algorithm and the KLT tracking algorithm is proposed.And in tracking experiments on target images in multiple scenes,the KLT tracking algorithm takes about 44.05%of the improved algorithm SUFREAK,which can effectively replace the image matching algorithm for target tracking,and improves the fluency of the mobile augmented reality system.Finally,a mobile augmented reality prototype system for campus building scene is developed based on Android.