Research And Implementation Of Cloud And Edge Collaborative Computing System In Web AR

Web AR has a broad application prospect as a universal and lightweight solution in the field of augmented reality.However,it still faces great challenges in practical applications.Web browsers are limited by their computing power to meet the computing and data intensive requirements of AR itself,and the physical distance and network distance between the central cloud computing data center and users are too far to meet the delay and bandwidth requirements of Web AR under the condition of balancing costs,resulting in a bad user experience.As a complementary technology of cloud computing,edge computing uses edge computing resources close to the data source to complete calculation and storage,which makes it better than traditional centralized cloud computing in terms of service response delay and bandwidth supply.Based on the augmented reality service platform of the research group,this thesis introduces edge computing on the basis of cloud computing and provides customers with Web AR solutions for advertising marketing and commercial promotion and other needs.The research work of this thesis is divided into two aspects.On the one hand,to meet the core AR computing requirements,combined with the relevant computer vision algorithms,some image computing tasks are designed and implemented in detail on the edge node and user browser.A real-time tracking algorithm combining target detection with optical flow method is proposed,which completes the landing of Web AR task with large computational complexity and high real-time requirements.On the other hand,in the face of some problems brought by the introduction of edge computing,based on the Kubernetes orchestration architecture,the deployment and orchestration of edge services and the management of edge nodes in the cloud are realized.Finally,the related experiments are designed to verify the overall process of Web AR.Thr-ough the analysis of the experimental data of other schemes,the introduction of edge computing can effectively reduce the delay of Web AR and optimize the bandwidth occupation.The superiority of cloud and edge collaborative computing in Web AR field is verified.and it has certain reference value for the landing of other similar emerging services.