Design And Implementation Of AR Mobile Learning Platform Based On AKAZE Image Matching

With the Internet as the carrier,mobile learning breaks the time-space limit through computer technology,makes the way of acquiring knowledge more convenient and promotes the sharing of educational resources.However,at present,mobile learning has a single form of expression and practice is difficult to carry out.The augmented reality technology displays the superposition of virtual 3d scene in the real world,but the real-time image matching performance of mobile terminal has a bottleneck.In this paper,augmented reality technology is combined with the MOOC mode to develop a mobile learning platform with AR interactive function.The main work is as follows:This paper analyzes the basic definition of image matching technology,and studies the image matching technology based on AKAZE algorithm.AKAZE algorithm constructs the nonlinear scale space,accelerates the nonlinear filtering through the Fast Explicit Diffusion,optimizes the image feature description process through the M-LDB descriptor,and greatly improves the real-time and stability of image matching.The experiment is carried out to compare the comprehensive performance of AKAZE algorithm with ORB algorithm and BRISK algorithm in image matching.The results show that AKAZE algorithm is more in line with the requirements of real-time image matching performance in augmented reality technology.Then the workflow of image matching scheme based on AKAZE algorithm is studied,and applied to the tracking and registration of augmented reality technology.The Unity engine is used to build the platform,and the OpenCV library is used to build the main program of image matching.Users can scan images with the camera of mobile intelligent devices,the main program take frames from the video clips,do image feature detection through AKAZE algorithm,and then through the 64 M-LDB descriptors to describe image features,do feature matching by the ratio+RANSAC algorithm.The 3D virtual scenes will be rendered on the images if the features are matched.The AR mobile learning platform is designed and implemented,which is divided into the course learning sub-platform and the AR interaction sub-platform according to its functions,including five functional modules of user information,course management,exercise examination,topic discussion and AR interaction.The AKAZE algorithm is introduced to realize the image matching scheme,and the parameters were adjusted to meet the performance requirements of real-time image matching in AR interaction module.The AR mobile learning platform combines augmented reality technology with mobile learning.Users can not only learn online,but also develop practical and hands-on abilities through AR interaction such as simulated dialogues,simulated experiments and interactive question-and-answer.The course content is displayed more effectively,the form of teaching is more comprehensive,and the knowledge dissemination is more efficient.