| The demand for remote meetings,immersive social interaction,and collaborative office work with multiple participants is increasing.Currently available video conferencing and remote work software cannot fully meet people’s needs.With the development of augmented reality technology and the concept of the metaverse,developing a shared,collaborative AR system with immersive experiences has become the target and vision of many governments and internet companies worldwide.Head-mounted AR devices provide the possibility for this vision,but traditional head-mounted AR devices are expensive and uncomfortable to wear.To address these issues,some manufacturers have introduced lightweight head-mounted AR devices,but they still have problems with insufficient computing power and poor rendering quality.Based on this background,this paper presents the following innovative works:(1)To solve the problem of weak computing power and poor rendering quality of lightweight head-mounted AR devices that cannot support heavy rendering tasks in shared collaborative systems,this paper proposes a multi-terminal,cloud-based AR technology solution.Based on cloud rendering technology,the heavy rendering tasks are completed by cloud servers,greatly reducing the hardware requirements of the terminal devices.The rendering results are then overlaid onto the real scene using visual positioning algorithms and virtual and real space coordinate transformation technology.Additionally,the solution proposes a method of generating 3D bounding boxes for 2D video frame target objects to achieve multi-user interaction under cloud rendering.(2)To address the problem of rendering frame rate and user interaction delay caused by the concentration of intensive computing on cloud servers as the number of users increases,this paper proposes and designs an end-cloud fusion rendering algorithm.The algorithm optimizes the rendering frame rate and interaction delay of the system and establishes an evaluation function based on the current number of users and the vertex and texture information of the rendering model.It uses a taboo search algorithm to adaptively schedule rendering tasks between cloud servers and terminal devices,improving the overall rendering frame rate,decreasing the interaction delay,and increasing the maximum number of users the system can accommodate.Based on these two works,this paper designs and implements a multiterminal,collaborative AR system and tests the system’s functionality and performance.The test results show that the multi-terminal,collaborative AR system can meet the basic requirements of multi-user collaboration in terms of functionality.In terms of performance,the system can support up to 15-18 terminal users for collaborative sharing with guaranteed real-time rendering frame rate and interaction delay. |