Research In Augmented Reality Tracking Registration Technique Based On Android Platform

Augmented Reality uses computer technique to put virtual information into the real world and combine the virtual objects with real environment to enhance people's experience and perception of real environment. In recent years, mobile Augmented Reality system gets more and more attention of the scholars because of its convenience, great interactivities and terrific user experience.Tracking registration technique is the key to mobile Augmented Reality system, which directly decides the system performance. The current mobile Augmented Reality tracking registration arithmetic has slow speed and low real-time. Aiming at the above problems, the essay mainly study on the design and implementation of the mobile Augmented Reality system based on the natural feature points, the main research contents as follows:1. The essay does research on the feature matching and tracking problem of the mobile Augmented Reality system, it puts up with a tracking registration method based on the ORB algorithm and L-K optical flow algorithm tracking technique on the mobile platform. ORB algorithm as a kind of binary feature description algorithm, which calculates quicker and has high efficiency, has been improved, with ORB algorithm to test and match the natural feature,it can shorten the time loss of feature extraction and matching stage and speed up the registration. Meanwhile, using L-K optical flow algorithm to track features'location in real time can adequately find out the correlation between the neighboring frames and improve algorithm's implementation efficiency.2. A mobile Augmented Reality system based on the above tracking registration algorithm has been designed and finished on the Android phone platform. The system uses the analytical feature matching and tracking module to gain the feature information, which is used to complete camera pose estimation; then, the effect to augment reality is completed with the OpenGL ES technique to put virtual model into the real environment to finish the fusion technology of combining the virtual and the real. Experiments show that the system the essay raises can not only satisfy mobile device requirements for real-time, but also suit different scales of scenes, rotating changes and partially occluded changes, which has great robustness.