Technology Research Of Real Time Augmented Reality System On Android Platform

With the development of science and technology,computer image fusion processing combined with video analysis,it is becoming more and more popular in various industries,However,as one of the new areas of development,Augmented reality technology has gradually received attention from all of life.It is one of the trends in the future.Augmented reality technology extracts video and real-time video analysis to detect the specific location of the target object,then the computer generated virtual image superimposed on the real scene of the target object,then it achieves the effect of the actual situation.Augmented reality is worthy of research and exploration for various industries in the field of people.For example,for exhibitions,meetings,etc.a virtual three-dimensional image shows a scene in the real world,it enhances the user experience of sensory perception,enterprises can also bring considerable economic benefits and reduce labor costs through this technology.Based on feature point extraction,the paper proposed and implemented a real-time matching on Android system of augmented Reality.The main research is based on detection and matching of the image feature point,three dimensional registration method and key issues by Android platform.The goal can complete the actual situation and the integration,then can bring the same experience effect on the Android platform.Firstly,the position and orientation of the camera is calculated,the introduction of computer vision under the relevant coordinates,perspective transformation,the camera internal control parameters and Virtual three dimensional registration method.The translation and rotation matrix are calculated by projection equation formula camera model and though using the generated virtual objects into the real scene,it becomes the actual situation intuitive fusion.Then the paper introduces the common feature points detection,and proposes a new algorithm,which is used to speed up the image feature points detection and matching,which is the foundation for the realization of Android.After the key frame and the reference image acquisition,the feature points and the description information of the key frame and the reference image are acquired.in order to reduce the computational efficiency at the same time,The error matching points are eliminated by RANSAC method.Finally,Android system is researched and solved,it obtains key frame to the c++local development under Android provided and uses dynamic link libraries,not only it solves the Android under image processing slow problems and satisfy users to access information in the outdoors demand problem and convenience on the Android platform.The proposed method can fast image matching,and can display real-time virtual objects,augmented reality effect can be achieved stable and accurate registration.Augmented Reality-based experimental results show that the case of the reference image partially occluded and angles of different scales on the Android,it accurate tracking and better image matching efficiency.