Research On Mobile Augmented Reality Using Hybrid Features

Augmented reality merges the virtual with the real to produce a new environment where digital and physical objects co-exist and interact in real time by computer technology. With the rapid development of intelligent mobile terminals, augmented reality has become a new research focus with more and more research and applications. Mobile augmented reality is an important branch of augment reality and its mobile terminals mainly have the characteristics of visual display, state awareness, network communication, scene interaction and so on. However, mobile terminals have the limitation of computational capacity compared with personal computers. Therefore, simple features are often employed in mobile augmented reality for camera tracking in real time. But it often leads to the unsteadiness of camera tracking.In this paper, a novel approach to real-time camera tracking in mobile augmented reality is proposed using hybrid features to solve this problem. By integrating feature points and lines as scene features, hybrid features are generated through the process of point/line feature extraction, optimization and fusion, and used to the real time estimation of camera parameters. A method for image feature optimization is proposed based on the scene structure analysis to meet the computational constraints of mobile terminals. In order to improve the stability of camera tracking, an iterative filtering method is put forward to choose a set of stable feature lines, and hybrid features are adaptively constructed based on the composition and geometry of scene features, which is used to improve the stability of camera tracking.In mobile augmented reality, the camera’s pose changes frequently when using a mobile terminal to obtain image sequences, which often leads to inconsistence of the image scene structure and the real scene structure and makes feature optimization fail based on the scene structure analysis. Meanwhile, the real scene changes with camera motion in a mobile terminal and the search window is fixed when feature matching, which leads to the same search range no matter how fast the real scene changes.Based on the characteristics of mobile terminals, a novel approach to feature tracking is proposed using multi-modal data in mobile augmented reality. The real scene structure can be restored from the image scene structure using multi-modal data. It makes feature optimization effective based on the scene structure analysis. Meanwhile, feature matching can be more accurate on a mobile platform using multi-modal data to generate an adaptive search window and it makes a better fusing effect of the virtual and the real.