Aiming at the problems of large data volume, long calculation time, and information feedback speed in traditional virtual augmented reality-based scenes, this paper constructed a virtual augmented reality scene based on 5G technology. First, based on the research of camera models and related calibration techniques, this article discusses the coordinate conversion and registration between virtual scenes and real scenes in augmented reality systems based on natural feature points. Secondly, the control mechanism of data synchronization and transmission are realized by using 5G communication technology. Then, through the approximate nearest neighbour search of the descriptor, the feature points are quickly matched. Finally, by tracking the QR-code logo black frame, the global holography matrix virtual real registration method relative to the initial image is used to calculate the camera’s posture and the fast tracking registration of the scene is realized. The experimental results prove that the scene constructed in this paper has the characteristics of real-time, wide field of view, high immersion and flexible expansion, which not only meets the requirements of matching accuracy, but also has the advantages of small calculation amount and fast calculation speed, and has good robustness and real-time.

针对传统虚拟增强现实场景中数据量大、计算时间长、信息反馈速度快等问题，构建了基于5G技术的虚拟增强现实场景。首先，本文基于对相机模型和相关标定技术的研究，讨论了基于自然特征点的增强现实系统中虚拟场景与真实场景的坐标转换和配准。其次，利用5G通信技术实现数据同步和传输的控制机制。然后，通过对描述符的近似最近邻搜索，快速匹配特征点。最后，通过跟踪二维码标志黑框，采用相对于初始图像的全局全息矩阵虚实配准方法计算相机的姿态，实现场景的快速跟踪配准。实验结果证明，本文构建的场景具有实时性、视野广、沉浸感强、扩展灵活等特点，既满足匹配精度要求，又具有计算量小、计算速度快，具有良好的鲁棒性和实时性。